WP4.3-CPK-D4.3.2-Medium Access Control For A Wireless Ad-Hoc Environment-13!10!2003-V1.0

INFORMATION SOCIETIES TECHNOLOGY (IST) PROGRAMME
PACWOMAN
Power Aware Communications for

Wireless OptiMised personal Area Network
Contract No IST-2001-34157
D 4.3.2 - Medium Access Control for Wireless Personal Area Networks

IST-2001-34157
Deliverable number: D 4.3.2
Title of Deliverable: Medium Access Control for Wireless Personal Area Networks
Contractual Date of Delivery to the CEC: August 31, 2003

Actual Date of Delivery to the CEC: October 14, 2003
Author(s): Andreas Hutter
John Spanoudakis
Ove Edfors
Carl Wijting
Participant(s): CSEM, AAU/CPK, Lund. Univ., Miltech
Workpackage: W4.3. MAC layer development, parameter
extraction and optimisation
Est. person months: 15
Security level: Pub*
Nature: Report
Version: 1.0
Total number of pages: 58
Abstract:
This deliverable presents the results of the studies undertaken in the context of the PACWOMAN project towards
MAC layer development, parameter extraction and optimisation. The deliverable investigates MAC mechanisms
for Wireless Personal Area Networks, for both High and Low Data Rate environment. The main part of the
deliverable focuses on the development of MAC mechanisms for ad-hoc networking and new solutions for
medium access control in the ad-hoc networking domain are proposed. The deliverable develops a simulation
model and approach to investigate the MAC protocols. Reviewing simulation platforms, discussing their suitability
and presenting simulation scenarios. An approach based on SDL is introduced, which will be used for functional
verification and an estimate of the implementation complexity of the proposed MAC extensions.
Keyword list: Medium Access Control, Carrier sensing, H-OFDM, Ultra-WideBand, Simulation techniques, SDL
 Copyrighted by the Participant(s)
Page 2 of 65
IST-2001-34157
DOCUMENT REVISION HISTORY
Version Date Author Summary of main changes

0.0 2003-05-13 Carl Wijting Initial Table of Contents (ToC)
0.1 2003-06-23 Carl Wijting Extension of ToC
0.2 2003-08-15 Andreas Hutter CSEM’s contribution
0.3 2003-09-10 John Spanoudakis Miltech’s contribution
0.4 2003-09-17 Carl Wijting CPK’s contribution +
first circulated complete draft
0.5 2003-09-23 Carl Wijting Final draft for approval PMB
0.6 2003-10-12 Carl Wijting Modified version after comments by PMB
1.0 2003-10-13 Ingrid Reynaert Approved
This document contains material, which is the copyright of certain PACWOMAN contractors, and may
not be reproduced or copied without permission.
The commercial use of any information contained in this document may require a license from the
proprietor of that information.
The contractors do not warrant that the information contained in the report is capable of use, or that
use of the information is free from risk, and accept no liability for loss or damage suffered by any
person using this information.
Page 3 of 65
IST-2001-34157
EXECUTIVE SUMMARY
Background
The IST PACWOMAN (Power Aware Communications for Wireless OptiMised personal Area Network)
project aims at performing the necessary research, development and optimisation on all OSI layers,
enabling the design of a low-cost, low-power, and flexible WPAN-based system.
This deliverable presents the results of the studies undertaken in the context of the PACWOMAN
project towards MAC layer development, parameter extraction and optimisation, which is a sub-work
package of Work package 4 on Radio/PHY/MAC development, parameter extraction and optimization.
Aim
The deliverable investigates MAC mechanisms for Wireless Personal Area Networks, for both High
and Low Data Rate environment. The main part of the deliverable focuses on the development of
MAC mechanisms for ad-hoc networking and new solutions for medium access control in the ad-hoc
networking domain are proposed. The deliverable develops a simulation model and approach to
investigate the MAC protocols. An approach based on SDL is introduced, which will be used for
functional verification and an estimate of the implementation complexity of the proposed MAC
extensions.
Conclusions
Medium Access Control in the Wireless PAN environment is divided into two fields, one focussing on
the medium and high data rate domain where ad-hoc routing techniques are applied and the other
focussing on the low data rate domain where typically master-slave communication is used. In the
MDR/HDR domain also multi-hop connections are considered.
For both these domains requirements and design constraints are considered and based upon this
design issues for the MAC layer are presented.
For the ad-hoc domain extensions to carrier-sensing techniques have been considered and a receiver
initiated MAC algorithm that applies energy pulses to jam the signal is proposed. An initial study on
link adaptation for UWB systems was performed, where the issues and parameters were identified to
design an adaptable MAC layer for UWB. This MAC can adjust the physical parameters to meet a
predetermined BER.
Several currently available simulation platforms were reviewed and it was concluded to use either
NS/2 or custom-made tools. NS/2 will be used for large system level evaluations, where the whole
protocol stack is simulated and the effects of multiple layers are studied. Custom-made tools are used
to test small algorithms and ideas, before a system level implementation is done.
A SDL specification of the MAC layer of IEEE 802.11 has been initiated to analyse the implementation
complexity of proposed extensions to CSMA MAC. This approach will be used to further analyse the
extensions to the carrier sensing mechanisms from an implementation point of view. In particular
focussing on power consumption and power saving.
Page 4 of 65
IST-2001-34157
CONTENT
DOCUMENT REVISION HISTORY........................................................................................3

EXECUTIVE SUMMARY........................................................................................................4
CONTENT.............................................................................................................................. 5
ABBREVIATIONS.................................................................................................................. 8
REFERENCES....................................................................................................................... 9
1 INTRODUCTION............................................................................................................ 11
2 MAC TECHNIQUES IN GENERAL................................................................................12
2.1 Basic Principles.........................................................................................................................................12
2.2 Architecture Design..................................................................................................................................13
2.3 Power-saving mechanisms.......................................................................................................................14
2.4 Relevant standards...................................................................................................................................15

2.4.1 Wireless LAN: IEEE 802.11..............................................................................................................15
2.4.1.1 The IEEE 802.11 Wireless LAN Architecture...............................................................................16
2.4.1.2 IEEE 802.11 Topologies.................................................................................................................16
2.4.1.3 IEEE 802.11 Media Access Control...............................................................................................17
2.4.2 Bluetooth.............................................................................................................................................18
2.4.2.1 Bluetooth Specifications.................................................................................................................19
2.4.2.2 Low power modes...........................................................................................................................20
2.4.3 IEEE 802.15.3.....................................................................................................................................20
2.4.4 IEEE 802.15.3a...................................................................................................................................21
2.4.5 IEEE 802.15.4.....................................................................................................................................21
3 MEDIUM ACCESS CONTROL FOR THE AD-HOC LINK.............................................24
3.1 Requirements and Constraints................................................................................................................24

3.1.1 Extensions to CSMA/CA....................................................................................................................24
3.1.2 MAC layer for H-OFDM....................................................................................................................26
3.2 Design Considerations..............................................................................................................................28

3.2.1 Extensions to CSMA..........................................................................................................................28
3.2.2 H-OFDM.............................................................................................................................................28
3.3 MAC alternatives......................................................................................................................................29
3.4 Implementation Proposal.........................................................................................................................30

3.4.1 Basic Protocol Operation....................................................................................................................30
3.4.2 Contention resolution..........................................................................................................................32
3.4.3 The choice of tsched...............................................................................................................................33
4 MEDIUM ACCESS CONTROL FOR THE LOW DATA RATE LINK..............................35
4.1 Requirements and Constraints................................................................................................................35
4.2 Design Considerations..............................................................................................................................36
Page 5 of 65
IST-2001-34157
4.3 MAC Alternatives.....................................................................................................................................37

5 SIMULATION TECHNIQUES.........................................................................................38
5.1 Review of possible simulation platforms................................................................................................38

5.1.1 OpNet..................................................................................................................................................38
5.1.2 Network Simulator 2...........................................................................................................................38
5.1.3 GloMoSim..........................................................................................................................................39
5.1.4 Ptolemy II...........................................................................................................................................39
5.1.5 Tailor-Made Tools..............................................................................................................................39
5.1.6 Discussion and Simulator Choice.......................................................................................................40
5.2 Simulation Methodology and Scenarios.................................................................................................40

5.2.1 Single hop...........................................................................................................................................40
5.2.2 Single hop with hidden nodes.............................................................................................................41
5.2.3 Line structure......................................................................................................................................41
5.2.4 Meshed structure.................................................................................................................................41
5.2.5 Simulation Parameters........................................................................................................................42
6 MAC LAYER FORMAL DESCRIPTION AND MODELING............................................43
6.1 Description of Approach..........................................................................................................................43

6.1.1 Objectives...........................................................................................................................................43
6.1.2 SDL for MAC-layer Formal Description and Early Verification.......................................................43
6.1.2.1 SDL Benefits...................................................................................................................................43
6.1.2.2 SDL Principles................................................................................................................................43
6.2 Methodology..............................................................................................................................................46
6.2.1 Requirements Capture.........................................................................................................................46
6.2.2 SDL Specifications Phase...................................................................................................................46
6.2.2.1 High-level MAC-layer Architecture...............................................................................................47
6.2.3 SDL Functional Verification..............................................................................................................48
6.2.4 Complexity Evaluation and Optimisation of the SDL Model............................................................48
7 CONCLUSIONS............................................................................................................. 50
APPENDIX A: MEDIUM ACCESS TECHNIQUES IN ULTRA-WIDEBAND AD HOC
NETWORKS......................................................................................................................... 51
1. INTRODUCTION.........................................................................................................51
2. ULTRA-WIDEBAND PHYSICAL LAYER (UWB PHY)..............................................52
3. MEDIUM ACCESS ISSUES FOR IR-UWB WPAN.....................................................53
4. SYSTEM DESIGN FOR INTERFERENCE MITIGATION............................................55
5. CONCLUSION............................................................................................................ 56
6. REFERENCES............................................................................................................ 57
Page 6 of 65
IST-2001-34157
Page 7 of 65
IST-2001-34157
ABBREVIATIONS
aT Advanced Terminal
ACK Acknowledgement
BB Black Burst
bT Basic Terminal
CAN Community Area Network
CAP Contention access period
CFP Contention free period
CSMA Carrier sense multiple access
CSMA/CA Carrier sense multiple access with collision avoidance
CSMA/CD Carrier sense multiple access with collision detection
CTS Clear-to-send
DCF Distributed Co-ordination Function
DIFS Distributed inter-frame space
FAMA Floor Acquisition Multiple Access
FFD Full function device
GTS Guaranteed time slot
H-OFDM Hybrid OFDM
IEEE Institute for Electrical and Electronic Engineers
LAN Local area network
LDR Low data rate
LL Link layer
MAC Medium access control
MDR Medium data rate
OFDM Orthogonal Frequency Division Multiplexing
OFDMA Orthogonal Frequency Division Multiple Access
PAN Personal Area Network
PCF Point Co-ordination Function
PHY Physical layer
PIFS PCF inter-frame space
POS Personal operating space
RFD Reduced function device
RTS Ready-to-sent
SIFS Short inter-frame space
TDD Time division duplex
UWB UltraWide Band
WAN Wide Area Network
WLAN Wireless local area network
WPAN Wireless personal area network
Page 8 of 65
IST-2001-34157
REFERENCES
[1] Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications, ‘The
Institute of Electrical and Electronics Engineers, IEEE 802.11, 1997.
[2] J. C. Haartsen, "The Bluetooth Radio System", IEEE Personal Communications, vol. 7, no. 1,
February 2000 pp. 28-36.
[3] Bluetooth SIG, www.bluetooth.com.
[4] IEEE Std. 802.15.4-2003, Wireless Medium Access Control (MAC) and Physical Layer (PHY)
specifications for Low Rate Wireless Personal Area Networks (LR-WPANs).
[5] E. Callaway, P. Gorday, L. Hester, J.A. Gutierrez, M. Naeve, B. Heile, V. Bahl, ‘Home networking
with IEEE 802.15.4: a developing standard for low-rate wireless personal area networks,’ IEEE
Commun. Mag., Aug. 2002.
[6] PACWOMAN Consortium, ‘State-of-the-Art of the WPAN networking paradigm’, Deliverable 5.1,
PACWOMAN IST-2001-34157, 2002.
[7] D. Porcino, W. Hirt, ‘Ultra-wideband radio technology: potential and challenges ahead,’ IEEE
Commun. Mag., July 2003.
[8] W. Ye, J. Heidermann, D. Estrin, ‘An energy-efficient MAC protocol for wireless sensor networks,’
USC/ISI Technical Report ISI-TR-543, Sept. 2001.
[9] S. Singh and C.S. Raghavendra, ‘PAMAS: Power aware multi-access protocol with signalling for
ad-hoc networks,’ ACM CCR, 28, 3, 5-26, July, 1998.
[10] A. El-Hoiydi, ‘Spatial TDMA and CSMA with Preamble Sampling Aloha with Preamble for Low
Power Ad Hoc Wireless Sensor Networks,’ Proc. IEEE Intl. Conf. Comp. and Commun. (ISCC
2002), Taormina, Italy, July 2002, pp. 685-692.
[11] Y-C Tseng, C-S Hsu, and T-Y Hsieh, ‘Power-Saving Protocols for IEEE 802.11-Based Multi-Hop
Ad Hoc Networks’, IEEE INFOCOM, 2002, pp. 200-209.
[12] PACWOMAN Consortium, ‘State-of-the-Art of the WPAN networking paradigm’, Deliverable 5.1,
PACWOMAN IST-2001-34157, 2002.
[13] PACWOMAN Consortium, 'Sytem Requirements and Analysis’, Deliverable 2.1, PACWOMAN
IST-2001-34157, 2002.
[14] PACWOMAN Consortium, 'Functional and Technical Specifications’, Deliverable 2.2,
PACWOMAN IST-2001-34157, 2002 (Internal Deliverable).
[15] J.L. Sobrinho, A.S. Krishnakumar, ‘Quality-of-Service in Ad Hoc Carrier Sense Multiple Access
Wireless Networks’, IEEE Journal on Selected Areas in Communications, Vol. 17, No 8, August
1999.
[16] C. Papachristou, F.N. Pavlidou, Collision-free operation in ad hoc carrier sense multiple access
wireless networks, IEEE Communications Letters, Vol. 6, No. 8 , Aug. 2002, pp. 352 –354.
[17] C.S. Wijting, R. Prasad, ‘Evaluation of Mobile Ad-hoc Network Techniques in cellular systems’,
IEEE conference on Vehicular Technology 2000 fall, Boston, U.S.A., 2000, pp. 1025-1029.
[18] J. Broch, D. A. Maltz, D. B. Johnson, Y. C. Hu, and J. Jetcheva. ‘A Performance Comparison of
Multi-Hop Wireless Ad Hoc Network Routing Protocols’. Proc. of the ACM/IEEE MobiCom,
October 1998.
[19] S. Ni, Y. Tseng, Y. Chen, and J. Sheu, The broadcast storm problem in a mobile ad hoc network,
ACM MOBICOM '99, August 1999.
[20] S. Xu, T. Saadawi, ‘Does the IEEE 802.11 MAC protocol work well in multihop wireless ad hoc
networks?’, IEEE Communications Magazine, Vol. 39 No. 6, June 2001, pp. 130 –137.
[21] E.M Royer, Sung-Ju Lee, C.E. Perkins, ’The effects of MAC protocols on ad hoc network
communication’, IEEE Wireless Communications and Networking Conference, 2000, Volume: 2 ,
September 2000, pp 543 –548.
[22] H. Yomo, P. Popovski, C.S. Wijting, I. Z. Kovács, N. Deblauwe, A. F. Baena, and R. Prasad,
`Medium Access Techniques in Ultra-Wideband Ad Hoc Networks´, ETAI 2003, Ohrid, Macedonia,
Sept. 2003.
[23] S.L. Wu, Y.C. Tseng, and J.P. Sheu, ‘Intelligent Medium Access for Mobile Ad Hoc Networks with
BusyTones and Power Control‘, IEEE Journal on Selected Areas in Communications, Vol 18, pp.
September 2000 1647-1657.
Page 9 of 65
IST-2001-34157
[24] R. Ramanathan, and R. Rosales-Hain, ‘Topology Control of Multihop Wireless Networks using
Transmit Power Adjustment’, IEEE INFORCOM, 2000, pp. 404-413.
[25] C. F. Huang, Y. C. Tseng, S. L. Wu, and J. P. Sheu, “Increasing the Throughput of Multihop
Packet Radio Networks with Power Adjustment,” International Conference on Computer,
Communication, and Networks, 2001.
[26] H. Woesner, J.P. Ebert, M. Schlager, and A. Wolisz, ‘Power-Saving Mechanisms in Emerging
Standards for Wireless LANs: The MAC Level Perspective’, IEEE Personal Communications, June
1998, pp. 40-48.
[27] J. Gomez, A. T. Campbell, M. Naghshineh, and C. Bisdikian, “A Distributed Contention Control
Mechanism for Power Saving in random access Ad Hoc Wireless Local Area Networks,” IEEE
International Workshop on Mobile Multimedia Communications, 1999, pp. 114-123.
[28] J. H. Ryu and D. H. Cho, “A New Routing Scheme Concerning Power Saving in Mobile Ad Hoc
Networks,” Proc. of IEEE International Conference on Communications, vol. 3, 2000, pp. 1719-
1722.
[29] J. H. Ryu, S. Song, and D. H. Cho, “A Power Saving Multicast Routing Scheme in 2-tir
Hierarchical Mobile Ad Hoc Networks,” IEEE Vehicular Technology Conference, vol. 4, pp. 1974-
1978, 2000.
[30] S. Singh, M. Woo, and C. S. Raghavendra, “Power Aware Routing m Mobile Ad Hoc Networks,”
International Conference on Mobile Computing and Networking, 1999, pp. 181-190.
[31] F.A. Tobagi and L. Kleinrock, ‘Packet Switching in Radio Channels: Part II – The Hidden Terminal
Problem in Carrier Sense Multiple Access Modes and the Busy-Tone Solution’, IEEE Transactions
on Communications, Vol 23, No 12, 1975.
[32] Zhenyu Tang and J.J. Garcia-Luna-Aceves, 'Collision-avoidance transmission scheduling for ad-
hoc networks’, IEEE International Conference on Communications, June 2000, pp. 1788 –1794.
[33] PACWOMAN Consortium, ‘Analysis and simulation of physical layer requirements.
Synchronisation, PAR reduction and resource assignment algorithms. RF specifications’,
Deliverable 4.2.3., PACWOMAN IST-2001-34157, 2002.
[34] M. Gerla; R. Bagrodia; Lixia Zhang; K. Tang, and Lan Wang, ‘TCP over wireless multi-hop
protocols: simulation and experiments’, IEEE International Conference on Communications, June
1999, pp.1089 –1094.
[35] J.J. Garcia-Luna-Aceves and A. Tzamaloukas, ``Receiver-Initiated Collision-Avoidance in
Wireless Networks,'' ACM Wireless Networks, Special Issue on Best Papers from ACM/IEEE
Mobicom 99, Vol. 8, pp. 249-263, 2002.
[36] IEEE Std. 802.15.1-2002, ‘Wireless MAC and PHY Specifications for Wireless Personal Area
Networks (WPANs™)’,
Page 10 of 65
IST-2001-34157
1 INTRODUCTION
Being able to communicate anywhere, anytime, and any device is a global trend in today’s
development of communication and computing systems. Ubiquitous computing is becoming possible
by the advance of wireless communication technology and the availability of many light-weight,
compact, portable computing devices. The PACWOMAN (Power Aware Communications for Wireless
OptiMised personal Area Network) project aims at performing the necessary research, development
and optimisation on all OSI layers, enabling the design of a low-cost, low-power, and flexible WPAN-
based system.
This deliverable presents the results of the studies undertaken in the context of the PACWOMAN
project towards MAC layer development, parameter extraction and optimisation. The deliverable
investigates MAC mechanisms for Wireless Personal Area Networks, for both High and Low Data
Rate environment. The main part of the deliverable focuses on the development of MAC mechanisms
for ad-hoc networking and new solutions for medium access control in the ad-hoc networking domain
are proposed. The deliverable develops a simulation model and approach to investigate the MAC
protocols. Reviewing simulation platforms, discussing their suitability and presenting simulation
scenarios. An approach based on SDL is introduced, which will be used for functional verification and
an estimate of the implementation complexity of the proposed MAC extensions.
The Deliverable is organised as follows, Section 2 introduces Medium Access Control in general and
discusses the research and design issues in the context of the PACWOMAN architecture, then in
Section 3 the particular issues of the ad-hoc link are addressed and a receiver-initiated extension of
the IEEE 802.11 MAC is proposed based on channel jamming with energy pulses. Section 4
investigates the requirements and design issues for Medium Access Control in the Low Data Rate
domain of the PACWOMAN architecture. In Section 5 some simulation platforms are reviewed that
can be used for numerical analysis of the proposals, further simulation scenarios and parameters are
considered. In Section 6 formal description and modelling techniques are presented and applied to
IEEE 802.11, which will be used to analyse the protocol optimisations that are being proposed. Finally
in Section 7 Conclusions are drawn.
Page 11 of 65
IST-2001-34157
2 MAC TECHNIQUES IN GENERAL
2.1 Basic Principles

The introduction of multi-hop links introduces new research challenges to the design of short-range
wireless networks. The first design problem is the inherent presence of the hidden terminal effect.
Since nodes make multi-hop networks by definition not all nodes are within radio range of each other
and thus carrier sensing mechanisms alone are not sufficient.
Also in higher layers problems can occur, multi-hop connections introduce the need for routing
protocols to determine to which neighbouring node the packets have to be forwarded [12]. These
protocols should be able to deal with the characteristics of ad-hoc networks as discussed in the
previous Section. Several routing protocols have been proposed in recent days under which the most
promising for practical application are: Dynamic Source Routing (DSR) and Ad-hoc On-Demand
Vector (AODV) routing, see for more details [17] [18].
TCP or UDP are typically applied as a transport layer protocol over these ad-hoc routing protocols.
The performance of TCP/IP over erroneous links can be unstable, due to lost acknowledgements and
time-outs. The time-outs occur due to route discoveries that are invoked by nodes that are not able to
contact their adjacent nodes (e.g. due to set Network Allocation Vectors). This effect is enhanced by
the slow-start mechanism of TCP.
Broadcasting is expected to be a common operation in ad-hoc networks, for example for paging, route
discovery, or other kinds of signalling information. A straight forward manner to implement
broadcasting is flooding, however if the flooding is done blindly it will result in many collisions since
nodes will receive and try to forward the data almost simultaneously, resulting in many colliding
channel accessing attempts. This effect is called the broadcast storm problem and was first addressed
in [19].
A particularly important issue to be investigated with respect to multi-hop ad-hoc networks is power-
awareness throughout the design. Here we will focus on the power-awareness of the MAC protocols.
The first method to reduce the power consumption of the device is the implementation of a spectrum
efficient protocol, limiting the amount of unnecessary retransmissions and collisions to the absolute
minimum. A second method is control of the transmission power, by reducing the transmission power
the interference is reduced and the throughput increased. However a minimum power is required to
maintain connectivity, between these two boundaries an optimal topology can be derived. A third
method is power-aware routing, where the paths are optimised according to a certain cost function,
e.g. battery lifetime, or minimum power usage in the path. Finally also low-power modes can be
applied, where for example the devices only wake up periodically [11].
The multi-hop connections might cause serious unfairness problems, differences in round-trip time
may cause TCP connections with a shorter round-trip time to starve out TCP connections with a larger
one [20]. But there are also unfairness issues arising from the MAC layer, in particular because the
sensing range is typically larger than the reception range. This means that a data flow that is being
transmitted starves out connections that are in it's sensing range, even though the flow is not received
there and would not interfere with an ongoing connection. Basically this is due to the exposed terminal
problem. Another example of MAC layer unfairness is the fact that the exponential back-off
mechanism favours the latest successful node, since the back-off window of all unsuccessful nodes is
increased while the window of the successful one remains the same [20].
Given all the previous constraints, the offering of Quality-of-Service, i.e. bounded service levels, is
particularly difficult. In this chapter MAC protocol extensions will be investigated and developed to
Page 12 of 65
IST-2001-34157
offer bounded service levels in a multi-hop environment, while reducing the number of collisions and
thus limiting the power consumption of the devices.
The Medium Access (MAC) Layer is responsible for the establishment and control of the physical
channels provided by the PHY layer. The efficiency of MAC layer mechanisms is especially important
for packet transmission since only a small overhead can be tolerated here. In the context of Wireless
PAN systems also power-awareness is a crucial requirement.
The MAC layer has two main functions, to coordinate the accessing of a link by the involved terminals
(multiplexing of the different users on the link) and secondly to reduce the interference between links
within the system and between different systems. Recently several groups are investigating MAC layer
performance in multi-hop networks [21] [34].
2.2 Architecture Design
A reference architecture for Short-range wireless systems has been derived. The architecture consists
of three main connection concepts: Personal Area Networks (PAN), Community Area Networks
(CAN), and Wide Area networks (WAN). Where a PAN provides connectivity in the Personal
Operating Space of a user, the CAN provides direct connectivity between these PANs and the WAN is
used if no direct communication is possible. In this case an alternative system such as GSM/GPRS or
UMTS is used for connectivity. Three different types of terminals are distinguished: Basic Terminals,
Advanced Terminals and Master Terminals. The basic terminals have a limited protocol stack and
communicate in a Master-Slave manner with the Master, Master terminals can communicate with
basic terminals, as well as with advanced terminals, in a Master-Slave and Ad-hoc manner
respectively. Advanced terminals can communicate with each other and with Master terminals in an
Ad-hoc manner.
Within the PACWOMAN architecture two domains have been defined; the Master-Slave domain and
the Ad-hoc domain. The Master-Slave domain deals with the communication within a Virtual Device
consisting of a Master and several Slaves, the Ad-hoc domain deals with the communication between
Master Devices and Advanced Terminals. These domains are also referred to as Low-Data Rate
(LDR) and High-Data Rate (HDR). Figure 2 .1 presents a high level overview of the main concepts in
the PACWOMAN architecture as defined in [13][14].
M
M
aT
Figure 2.1 Overview of the high level PACWOMAN architecture.
Besides the HDR domain, also a Medium Data Rate (MDR) domain has been defined, basically here
Bit Rate has been exchanged for transmission range as compared to the HDR domain. Table 2 .1
summarises the main system requirement of the different domains as defined for the PACWOMAN
architecture in D2.2 for each of the different domains. Additionally, the maximum number of hops for
the MDR and HDR domain is set to 10. The LDR environment applies a star topology, whereas the
MDR and HDR use a meshed topology.
Page 13 of 65
IST-2001-34157
Table 2.1 Characteristics for the different domains: Low, Medium, and High Data Rate.
LDR MDR HDR
Data Rate 10 bps – 10 kbps 10 kbs – 1 Mbps > 1Mbps
Required BER 10-3 - 10-7 10-3 - 10-7 10-3 - 10-7
Maximum
1m – 5m < 20 m < 10 m
distance
Autonomy (power hours / days
months / years days / months
consumption) (main supply)
Within the PACWOMAN architecture different technologies are applied for these different links. This is
one of the characteristics of WPAN systems, heterogeneity of the access techniques. Within work
package 4 two new access methods for the physical layer are being investigated, Hybrid-OFDM (H-
OFDM) and Ultra Wideband (UWB) for the ad-hoc and master-slave domain respectively (see
deliverables 4.2.1, 4.2.2 and 4.2.3 for more details). In the demonstrator the concept will be prototyped
using Bluetooth and IEEE 802.11 WLAN.
Figure 2 .2 depicts the protocol stacks for both the basic and advanced terminal, with the different
possible physical layers. The basic terminal has possibly an even more limited protocol stacks, there
is no need for a simple sensor to run a full TCP stack.
NET (TCP/IP) NET (TCP/IP)
LL (IEEE 802.3) LL (IEEE 802.3)
MAC MAC (TDMA, CSMA…)
UWB BT IEEE 802.11 BT H-OFDM
Basic Terminal (bT) Advanced Terminal (aT)

Figure 2.2 The different PHY layers in within the PACWOMAN Project.
In this deliverable the requirement and design issues are addressed for the MAC layer of both the
basic and advanced terminal. One of the main requirements taken into account is need for power
saving mechanisms.
2.3 Power-saving mechanisms
A critical resource for all kinds of portable devices when operated on batteries is power. Without
battery power those devices cannot function, therefore power saving mechanisms are very important.
In addition it is expected that the efficiency increase of batteries will not be as fast as the increase of
computational complexity of computing devices.
Generally power-saving mechanisms for short-range wireless systems can be classified as follows
[11]:
Transmission Power Control

In wireless communication, transmission power has strong impact on bit error rate, transmission rate,
and inter-radio interference. These are typically contradicting factors. In [23], power control is adopted
to reduce interference and improve throughput on the MAC layer. How to determine transmission
power of each mobile host so as to determine the best network topology, or known as topology
control, is addressed in for example [24]. How to increase network throughput by power adjustment for
packet radio networks is addressed in [25].
Page 14 of 65
IST-2001-34157
Power-Aware Routing
Power-aware routing protocols have been proposed based on various power cost functions. In [27],
when a mobile host’s battery level is below a certain threshold, it will not forward packets for other
hosts. In [30], five different metrics based on battery power consumption are proposed. Reference [11]
considers both hosts’ lifetime and a distance power metric. A hybrid environment consisting of battery-
powered and outlet-plugged hosts is considered in [28]. Two distributed heuristic clustering
approaches for multicasting are proposed in [29] to minimizing the transmission power.
Low-Power Mode
More and more wireless devices can support low-power sleep modes. IEEE 802.11 [1] has a power-
saving mode in which a radio only needs to be awake periodically. HiperLAN allows a mobile host in
power-saving mode to define its own active period. An active host may save powers by turning off its
equalizer according to the transmission bit rate. Comparisons are presented in [26] to study the
power-saving mechanisms of IEEE 802.11 and HIPERLAN in ad hoc networks. Bluetooth [2] provides
three different low-power modes: snif, hold, and park. Another method for saving power (for not time
critical applications) in a mobile system is to hold transmission until the channel is favourable (i.e.
strong).
2.4 Relevant standards
In this Section we briefly summarise the main features of already today existing standards for short-
range wireless communication systems. We introduce IEEE 802.11 Wireless LAN, Bluetooth and
IEEE 802.15.x Wireless PAN
2.4.1 Wireless LAN: IEEE 802.11
In 1997 the IEEE adopted IEEE Std. 802.11-1997, the first wireless LAN (WLAN) standard. This
standard defines the media access control (MAC) and physical (PHY) layers for a LAN with wireless
connectivity. It addresses local area networking where the connected devices communicate over the
air to other devices that are within close proximity to each other. This Chapter provides an overview of
the 802.11 architecture and the different topologies incorporated. The standard is similar in many
respects to the IEEE 802.3 Ethernet standard. Specifically, the 802.11 standard addresses (see also ):
 Functions required for an 802.11 compliant device to operate either in a peer-to-peer fashion
or integrated with an existing wired LAN.
 Operation of the 802.11 device within possibly overlapping 802.11 wireless LANs and the
mobility of this device between multiple wireless LANs.
 MAC level access control and data delivery services to allow upper layers of 802.11 networks
to function transparently.
 Several physical layer signalling techniques and interfaces.
 Privacy and security of user data being transferred over the wireless media.
Station Management
LLC
OSI Layer 2 MAC MAC Management
(data link) PLCP
OSI Layer 1 PMD
(physical)
PHY Management
Figure 2.3 IEEE 802.11 standards mapped to the OSI reference model.
Page 15 of 65
IST-2001-34157
2.4.1.1 The IEEE 802.11 Wireless LAN Architecture
The IEEE 802.11 architecture is comprised of several components and services that interact to
provide station mobility transparent to the higher layers of the network stack. The basic architecture is
depicted in Figure 2 .4. Several entities can be distinguished, namely the Wireless Station or Mobile
Host, the Basic Service Set, the Distribution System and the Access Point
Wireless LAN Station: The station (or Mobile Host, MH) is the most basic component of the wireless
network. A station is any device that contains the functionality of the 802.11 protocol, that being MAC,
PHY, and a connection to the wireless media. Typically the 802.11 functions are implemented in the
hardware and software of a Network Interface Card (NIC). A station could be a laptop PC, handheld
device, or an Access Point. Stations may be mobile, portable, or stationary and all stations support the
802.11 station services of authentication, de-authentication, privacy, and data delivery.
Basic Service Set (BSS): as the basic building block of an 802.11 wireless LAN. The BSS consists of
a group of any number of stations. The BSS is not a very interesting topic until we take the topology of
the WLAN into consideration.
Distribution System (DS): is the means by which an access point communicates with another access
point to exchange frames for stations in their respective BSSs, forward frames to follow mobile
stations as they move from one BSS to another, and exchange frames with a wired network.
As IEEE 802.11 describes it, the distribution system is not necessarily a network nor does the
standard place any restrictions on how the distribution system is implemented, only on the services it
must provide. Thus the distribution system may be a wired network like 803.2 or a special purpose box
that interconnects the access points and provides the required distribution services.
Access Point: is a fixed entity in the network that connects the mobile hosts with the Distribution
System.
BSS
MH
MH AP DistributionSystem GW 802
C.xL
loudAN
MH AP
BSS
MH
Figure 2.4 The basic Architecture of an IEEE 802.11 Wireless LAN.
The coverage area of the wireless LAN can be extended via an Extended Service Set (ESS). An
extended service set is a set of infrastructure BSS’s, where the access points communicate amongst
themselves to forward traffic from one BSS to another to facilitate movement of stations between
BSS’s.
2.4.1.2 IEEE 802.11 Topologies

The 802.11 standard describes two modes of deployment for the WLAN, the infrastructure mode and
the ad hoc mode.
Page 16 of 65
IST-2001-34157
The most basic wireless LAN topology is a set of stations, which have recognized each other and are
connected via the wireless media in a peer-to-peer fashion. This form of network topology is referred
to as an Independent Basic Service Set (IBSS) or an Ad-hoc network.
In an IBSS, the mobile stations communicate directly with each other. Every mobile station may not be
able to communicate with every other station due to the range limitations. There are no relay functions
in an IBSS therefore all stations need to be within range of each other and communicate directly.
Distribution System
MH AP
MH
MH
MH
MH
MH
(a) (b)
Figure 2.5 The ad-hoc mode (a) and the infrastructure mode (b) of IEEE 802.11.
An Infrastructure Basic Service Set is a BSS with a component called an Access Point (AP). The
access point provides a local relay function for the BSS. All stations in the BSS communicate with the
access point and no longer communicate directly. All frames are relayed between stations by the
access point. This local relay function effectively doubles the range of the IBSS.
2.4.1.3 IEEE 802.11 Media Access Control

The MAC layer has to fulfil several tasks. First of all it controls of course the access to the channel, but
besides that it can also offer support for roaming, power conservation and authentication. In basic
mode the system offers asynchronous data services and an optional mode exists for offering time-
bounded services. Depending on the applied topology, the modes can be used. In the infrastructure
mode both options are available, whereas in the ad hoc mode only the asynchronous data services
are available.
The following three basic access schemes have been defined for IEEE 802.11:
1. The mandatory basic access method based on CSMA/CA
2. An optional method avoiding the hidden terminal problem
3. A contention free polling method for time-bounded service
The first two methods are implemented as the Distributed Coordination Function, the third method is
referred to as Point Coordination Function. These MAC mechanisms are referred to as Distributed
Foundation Wireless Medium Access Control (DFWMAC).
Several parameters have been defined for implementing the access methods (see Figure 2 .6).
DIFS DIFS
PIFS
Medium busy SIFS Contention Next Frame
t
Figure 2.6 Inter Frame Spacing for the IEEE 802.11 Medium Access.
Page 17 of 65
IST-2001-34157
DCF Inter-Frame Spacing this is the longest waiting time available and thus grants the lowest priority
for accessing the medium, after this time-out CSMA/CA is used for accessing the channel.
PCF inter-Frame Spacing this is the medium waiting time and it is used for the time-bounded
services, which thus have priority over the asynchronous CSMA/CA services. Thus after the IFS
expired the system can move into polling mode.
Short Inter-Frame Spacing this is the shortest waiting time and thus it has the highest priority. This
waiting time is used for control messages, e.g. acknowledgements, polling responses, RTS / CTS
messages etc.
2.4.2 Bluetooth
The Bluetooth specification [3] has been produced by the Bluetooth Special Interest Group (SIG). The
Bluetooth SIG is an industry group consisting of leading manufacturers in telecommunications and
computing. The Bluetooth technology was initiated as a means to connect cell phones to laptop
computers, but it is expected that it will gradually replace cables and infrared as the primary means of
wirelessly exchanging information between devices.
Bluetooth is also set to become a link between a multitude of future devices. Examples include using
Bluetooth as the link between personal e-payment devices and point-of-sale terminals. There are even
proposals for Bluetooth motion-sensing rings that would allow fingers to act as electronic pointing
devices. These examples affirm the fact that Bluetooth is the first technology that embodies the PAN
concept. However, Bluetooth will only deliver best-case real-world data speeds of 720 kbps, which
would be insufficient for a number of applications, such as real-time video. This is one more argument
for further extensive research in personal area networking solutions.
For the scenarios envisioned by Bluetooth, it is highly likely that a large number of ad hoc connections
will coexist in the same area without any mutual coordination. This is different from ad hoc scenarios
considered in the past, where ad hoc connectivity focused on providing a single (or very few)
network(s) between the units in range, as it is the case with the wireless LANs. For the Bluetooth
applications, typically many independent networks overlap in the same area. This is referred as a
scatter ad hoc environment [2]. The difference between a conventional cellular environment, a
conventional ad hoc environment, and a scatter ad hoc environment is illustrated in Figure 2 .7.
(a (b (c
) ) )
Figure 2.7 Example topologies for: (a) cellular radio systems with squares representing base stations;
(b) conventional ad hoc systems; (c) scatter ad hoc environment.
Bluetooth is a de facto standard for present WPAN technology. In addition, the functional requirements
of Bluetooth specification are very similar to the ones proposed of IEEE 802.15 (more precisely
802.15.1). A large portion of Bluetooth will be a key ingredient of the potential 802.15.1 standard. The
usage of Bluetooth is expected to be so widespread, that each future PAN technology must consider
the Bluetooth compatibility.
Page 18 of 65
IST-2001-34157
2.4.2.1 Bluetooth Specifications
The Bluetooth specification is divided into two parts: the Core and the Profiles. The Core portion
specifies the architectural details of the Bluetooth system, hardware components, protocols and
interoperability with different legacy communication protocols. The Profiles portion specifies the
protocols and procedures required for different Bluetooth applications. We will examine the most
important aspects of the Bluetooth specification [2] [3] [36].
The Bluetooth system operates in the unlicensed ISM frequency band at 2.4 GHz. Regarding the
choice of the multiple access technique, CDMA offers the best properties for ad hoc radio systems
intended to operate in unlicensed spectrum. However, direct sequence (DS)-CDMA is less attractive
for the short-range wireless connectivity due to the excessive usage of battery power for DSP and
power control.
Bluetooth is based on frequency hopping (FH) – CDMA. In the ISM band, a set of 79 hop carriers has
been defined at a 1 MHz spacing 1. The hopping rate is 1600 hops per second, which provides good
immunity against the sources of interference in the 2.4 GHz band. The applied Gaussian Frequency
Shift Keying (GFSK) modulation achieves link speed of 1 Mbps.
A channel is defined by a particular pseudorandom FH sequence. Two or more (up to eight) active
Bluetooth units that share the same channel form a piconet. Among all units participating in a piconet,
only one is distinguished as a master, and the others are referred as slaves. The master determines
the particular hopping sequence associated with the piconet and the phase of the FH sequence is
defined by the master’s native clock.
The Bluetooth system provides full duplex transmission based on slotted time-division duplex (TDD).
The duration of a single hop is 625 s and the communication within a single slot is performed on one
frequency. Since the nominal hopping rate is 1600 hops per second, the working frequency changes
in every slot (exception are the multislot packets, explained further) There is no direct transmission
between slaves in a Bluetooth piconet, only from master to slave and vice versa.
Communication in piconet is organized so that the master polls each slave according to a polling
scheme. A slave is only allowed to transmit after having been polled by the master. The slave will start
its transmission in the slave-to-master time slot immediately after it has received a packet from the
master. The master may or may not include data in the packet used to poll a slave. However it is
possible to send packets that cover multiple slots. The multislot packets may be either 3 or 5 slots long
and during the multislot packet the frequency is not changed. The frequency applied after a multislot
packet is the same as the frequency that would be applied in the observed slot if there was not a
multislot packet. In general, the slots in which the master starts the transmission do not change their
ordering parity (the same is valid for the slaves). Since the master schedules the traffic in both the
uplink and downlink, intelligent scheduling algorithms have to be used that take into account the slave
characteristics.
The number of units that can participate in a piconet is deliberately limited to 8 (7 active slaves) in
order to keep high-capacity links between all the units. If the master needs to communicate with more
than seven devices, it can do so by first instructing active slave devices to switch to low-power park
mode and then inviting other parked slaves to become active in the piconet. However, there may be a
need for simultaneous connection of more than 8 devices. Thus, two or more interconnected piconets
form a scatternet. A Bluetooth unit can simultaneously be a slave member of multiple piconets, but
only a master in one. In principle, it is best for the bridge node to be a slave in both piconets, since
otherwise, the throughput within the piconet where the bridge node is a master is degraded. The
bridge node schedules its presence in the piconets in which it participates in a time-division manner.
Therefore, the global connectivity and throughput in the scatternet depend on the individual schedules
of the bridge nodes, as well as the schedules determined by the piconet masters.
1
At the time the specification was produced, for France, Spain and Japan a reduced set of 23 hop carriers has been defined at
a 1 MHz carrier spacing
Page 19 of 65
IST-2001-34157
The scatternet functionality is important to allow a flexible forming of ad hoc PANs. Scatternets are an
instance of the ad hoc network concept. When the interpiconet nodes forward packets between
piconets, Bluetooth ad hoc PANs belong to the class of multihop ad hoc networks. These scatternet-
based PANs may be used when information needs to be spread widely among the PANs residing
within reach through “reasonable” number of radio hops. Here, the reasonable number of hops
depends on the type of information in terms of required data rate and end-to-end delay.
Bluetooth uses a procedure known as inquiry for discovering other devices; it uses paging to
subsequently establish connections with them. The inquiry and paging are conceptually simple
operations, but the frequency-hopping nature of the physical layer makes the low-level details quite
complex. Two nodes cannot exchange messages until one of them becomes a master, thus
determining the frequency hopping sequence that will be used in the piconet. During inquiry, both
nodes (one is the listener and the other is the sender) hop using the same sequence; but the sender
hops faster than the listener, transmitting a signal on each channel and listening between
transmissions for an answer. When more than one listener is present, their replies may collide. To
avoid the collision, listeners defer their replies until expiration of a random backoff timer. Eventually
the sender device collects some basic information from the listeners, such as the device address and
the clock offsets. This information is subsequently used to page the selected listener device.
2.4.2.2 Low power modes
Bluetooth offers different low-power modes for improving battery life. Piconets are formed on demand
when communication among devices is ready to take place. At all other times, devices can be either
turned off or programmed to wake up periodically to send or receive inquiry messages. It is possible to
switch a slave into a low-power mode whereby it sleeps most of the time and wakes up only
periodically. Three types of low-power modes have been defined:
 Hold mode is used when a device should be put to sleep for a specified length of time. As
described earlier, the master can put all its slaves in the hold mode to suspend activity in the
current piconet while it searches for new members and invites them to join.
 Sniff mode is used to put a slave in a low-duty cycle mode, whereby it wakes up periodically
to communicate with the master.
 Park mode is similar to the sniff mode, but it is used to stay synchronized with the master
without being an active member of the piconet. The park mode enables the master to admit up
to 255 slaves in its piconet.
2.4.3 IEEE 802.15.3
The IEEE P802.15.3 high rate task group (TG3) for wireless personal area networks is chartered to
draft and publish a new standard for high-rate (20Mbit/s or greater) WPANs. The task group defined
the PHY and MAC specifications for high data rate wireless connectivity with fixed, portable and
moving devices within or entering a Personal Operating Space (POS). A goal of the high rate WPAN
task group was to achieve a level of interoperability or coexistence with other 802.15 Task Groups. It
is also the intent of this project to work toward a level of coexistence with other wireless devices in
conjunction with coexistence task groups such as 802.15.2. Currently the draft standard entitled
Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specs for High Rate Wireless
Personal Area Networks (WPAN) has been completed. The main features and characteristics of the
draft standard are:
- Data Rates: 11, 22, 33, 44, & 55 Mbps.

- Quality of Service isochronous protocol
- Ad-hoc peer-to-peer networking
- Security
Page 20 of 65
IST-2001-34157
- Low power consumption

- Low cost
- Designed to meet the demanding requirements of portable consumer imaging and multimedia
applications.
The draft standard is now in balloting phase. More details and updates on the balloting results can be
found on the task group site located at http://www.ieee802.org/15/pub/TG3.html.
2.4.4 IEEE 802.15.3a
The IEEE 802.15 high rate alternative PHY task group (TG3a) for wireless personal area networks is
working to define a project to provide a higher speed PHY enhancement amendment to 802.15.3 for
applications that involve imaging and multimedia. Out of initially 32 proposals for the alternative PHY
layer for the IEEE 802.15.3 high rate WPAN that have been presented in March 2003, 10 proposals
have been withdrawn and 17 proposals have been merged such that as of May 2003 only 6 proposal
remained. All these proposals are based on PHY layers that mostly operate with a channel bandwidth
of 500 MHz, such that the IEEE 802.15.3a can be qualified as an UWB-based system. The PHY layer
option presented in the different proposals include pulse-based modulation schemes such as pulse
position modulation (proposal no. 4) and soft spectrum adaptation (proposal no. 5) as well as
continuous modulation schemes such as DS-CDMA (proposal no. 1), direct sequence spread
spectrum (proposal no. 2), differential QPSK (proposal no. 3), and OFDM (proposal no. 6). The latter,
the so-called multiband OFDM proposal, is the merged proposal, which is likely to play an important
role due to its large support from industry. Note that these different proposals also provide a
description of the MAC layer. Details can be found on the TG3a home page located at
http://grouper.ieee.org/groups/802/15/pub/TG3a_CFP.html.
2.4.5 IEEE 802.15.4
The IEEE standard 802.15.4 [4] specifies the medium access control and the physical layer for low
rate wireless personal area networks (see also [5]). A low rate WPAN is a simple, low-cost
communication network that allows wireless connectivity in applications with limited power and relaxed
throughput requirements. The main objectives of a low rate WPAN are ease of installation, reliable
data transfer, short-range operation, extremely low-cost, and a reasonable battery life, while
maintaining a simple and flexible protocol. Some of the most important characteristics of a low rate
WPAN specified in the IEEE 802.15.4 standard are:
- Over air data rates of 250 kb/s, 40 kb/s, and 20 kb/s.

- Star or Peer-to-Peer operation.
- Allocation of guaranteed time slots.
- CSMA/CA channel access.
- Fully acknowledged protocol for transfer reliability.
- Low power consumption.
- Energy detection.
- Link quality indication.
- 16 channels in the 2450 MHz band, 10 channels in the 915MHz band, and 1 channel in the
868MHz band.
There are two different device types that can participate in an low rate WPAN network, a full function
device (FFD) and a reduced function device (RFD). The FFD can operate in three modes serving
either as a PAN coordinator, a coordinator or a device. An FFD can talk to RFDs or other FFDs, while
an RFD can only talk to an FFD. An RFD is intended for applications that are extremely simple, such
as a light switch or a passive infrared sensor, they do not have the need to send large amounts of data
and may only associate with a single FFD at a time. Consequently, the RFD can be implemented
using minimal resources and memory capacity.
Page 21 of 65
IST-2001-34157
A system conforming to IEEE 802.15.4 consists of several components, the most basic being the
device. A device can be an RFD or an FFD. Two or more devices within a personal operating space
(POS) communicating on the same physical channel constitute a WPAN. However, a network shall
include at least one FFD, operating as the PAN coordinator. An IEEE 802.15.4 type network is part of
the WPAN family of standards though the coverage of a low rate WPAN may extend beyond the POS,
which typically defines the WPAN. A well-defined coverage area does not exist for wireless media
since propagation characteristics are dynamic and uncertain.
The MAC sublayer provides two services; these are the MAC data service and the MAC management
service interfacing to the MAC sublayer management entity service access point. The MAC data
service enables the transmission and reception of MAC protocol data units across the PHY data
service. The features of the MAC sublayer are beacon management, channel access, guaranteed time
slot management, frame validation, acknowledged frame delivery, association and disassociation. In
addition, the MAC sublayer provides hooks for implementing application appropriate security
mechanisms.
The low rate WPAN standard allows the optional use of a superframe structure. The coordinator
defines the format of the superframe. The superframe is bounded by network beacons, is sent by the
coordinator and is divided into 16 equally sized slots. The beacon frame is transmitted in the first slot
of each superframe. If a coordinator does not wish to use a superframe structure it may turn off the
beacon transmissions. The beacons are used to synchronize the attached devices, to identify the
PAN, and to describe the structure of the superframes. Any device wishing to communicate during the
contention access period (CAP) between two beacons shall compete with other devices using a
slotted CSMA/CA mechanism. All transactions shall be completed by the time of the next network
beacon.
The superframe can have an active and an inactive portion. During the inactive portion the coordinator
shall not interact with its PAN and may enter a low power mode. For low latency applications or
applications requiring specific data bandwidth, the PAN coordinator may dedicate portions of the
active superframe to that application. These portions are called guaranteed time slots (GTSs). The
guaranteed time slots comprise the contention free period (CFP), which always appears at the end of
the active superframe starting at a slot boundary immediately following the CAP, as shown in Figure 2
.8. The PAN coordinator may allocate up to seven of these GTSs and a GTS may occupy more than
one slot period. However, a sufficient portion of the CAP shall remain for contention-based access of
other networked devices or new devices wishing to join the network. All contention-based transactions
shall be complete before the CFP begins. Also each device transmitting in a GTS shall ensure that its
transaction is complete before the time of the next GTS or the end of the CFP.
superframe beacon
contention contention
access period free period
Figure 2.8 The IEEE 802.15.4 WPAN superframe structure.
Page 22 of 65
IST-2001-34157
There are three types of data transfer transactions. The first one is the data transfer to a coordinator in
which a device transmits the data. The second transaction is the data transfer from a coordinator in
which the device receives the data. The third transaction is the data transfer between two peer
devices. In star topology only two of these transactions are used, since data may be exchanged only
between the coordinator and a device. In a peer-to-peer topology data may be exchanged between
any two devices on the network consequently all three transactions may be used in this topology. The
mechanisms for each transfer type depend on whether the network supports the transmission of
beacons. A beacon-enabled network is used for supporting low latency devices such as personal
computer peripherals. If the network does not need to support such devices, it can elect not to use the
beacon for normal transfers. However, the beacon is still required for network association purposes.
Page 23 of 65
IST-2001-34157
3 MEDIUM ACCESS CONTROL FOR THE AD-HOC LINK
This Section discusses the medium access control layer for the High Data Rate (HDR) or Ad-hoc
domain of the PACWOMAN Architecture. Within the PACWOMAN Architecture for this domain the H-
OFDM physical layer is being designed and analyzed for its suitability. Also the use of Wireless LAN
(IEEE 802.11) is investigated by means of demonstration. Therefore in this Section we will focus on
these two types of systems. In particular optimizations and extensions to the IEEE 802.11 MAC layer
will be considered, as well as design issues for the H-OFDM MAC layer.
First briefly the requirements for the HDR domain are reviewed and from this design constraints are
derived. After this design considerations are discussed, as well as suitability of techniques for the ad-
hoc domain and required modifications and adaptations. For the ad-hoc domain from several
proposed extensions to the IEEE 802.11 MAC one idea is designed in more detail: A receiver-initiated
scheme based on channel jamming with energy pulses.
3.1 Requirements and Constraints

The HDR link operates between different Personal Area Networks and possibly between several high
data rate devices within the PAN. The maximum distance between the individual devices is 10m,
however a larger range is possible by using multiple hops to transmit packets, which means that
intermediate nodes are used to forward the packets to the final destination.
Typical applications are (see also D2.1 and D2.2) tele monitoring, business applications with moderate
bit rate requirements and interactive video services or time-critical computer data transfers with high
bit rate requirements. These services require a bandwidth of over 1 Mbps with bit-error-ratios between
10-3 to 10-7. The delay requirements are set to up to 0.5s for the MDR domain and in the range of 10-
50ms for the HDR domain.
General requirements that apply to the medium access control in the ad-hoc mode are
 Energy constrain operation is of the utmost importance in the PACWOMAN context. The MAC
schemes should meet the power requirements of the Advanced terminals.
 Collisions should be limited, to prevent unnecessary transmissions. This way transmission
power is saved, especially under high load conditions. Here it is assumed that the transmit
power associated with packet transmission dominates the power consumption (over other
sources of power consumption such as the signal processing and channel sensing power).
 Increased efficiency, overhead reduction. By preventing unnecessary control messages from
being exchanged, both the protocols are more efficient and transmission power is saved.
 Flexibility, fast changing topologies, caused by new nodes arriving and others leaving the
network. Meaning that the developed MAC mechanisms should be able to cope with these
fast changing topologies and be able to quickly adapt to the new situation.
 Scalability, in the ad-hoc network a growing number of nodes can be incorporated (at the
moment ranging up to 10 hops).
 Co-existence of links (inter link interference), limiting the interference between links so that the
spectrum can be used efficiently
For the ad-hoc domain two types of techniques are being considered, carrier sensing and the Hybrid-
OFDM, the following paragraphs will review their basic properties, which are needed to derive the
design criteria.
3.1.1 Extensions to CSMA/CA
Typically Carrier Sense networks are designed such that there is a difference between the receiving
and the sensing range. Meaning that the threshold for sensing the channel is typically set (much)
Page 24 of 65
IST-2001-34157
lower than the power required for receiving a frame correctly. Due to these different ranges, a wireless
multi-hop network can be described by defining three types of links,
 Communication links, defined as links between two nodes in transmission range that have
packets to be exchanged between them.
 Interfering links, defined as links between two nodes in transmission range that have packets
to send to different destinations, but in overlapping time instances. These packets are said to
have collided
 Sensing links, defined as links between two nodes that if one of the nodes starts a
transmission the wireless link is sensed busy by the other node.
These three types of links can be described by connectivity matrices G: GC = (N, LC) for the matrix with
communication links, GI = (N, LI) for the interfering ones and finally GS = (N, LS) for the graph with the
sensing links, where N is the nodes in the system and LC, LI, and LS are the links (edges) that
communicate, interfere and allow sensing. The communication graph is a directed graph, the
interference and sensing graphs are not, the effect is reciprocal. GI is a spanning subtree of GS, LI  LS
[15].
Each link between nodes i and j has a propagation or path delay ij associated with it. Because of
these delays, the channel is not immediately sensed busy the moment the transmissions start.
In [15] Sobrinho proposes and investigates the basic Black Burst scheme. The Black Burst
mechanism is an extension that can be applied to CSMA system. Applying this mechanism, users
compete for the common radio channel by means of pulses of energy, referred to as Black Bursts
(BB). The length of these bursts depends on the time that users have been waiting for transmission as
of their scheduled moment of transmission. The algorithm is distributed and completely based on
channel sensing. It distinguishes two classes of users, one priority class and one best-effort class.
The protocol distinguished two different events, (i) obtaining the channel either as a best-effort user or
a non-scheduled priority user and (ii) obtaining the channel as a scheduled priority user. In the fist
case normal CSMA/CA is applied and in case of a priority class user the following transmission are
scheduled, in case of best-effort nothing further is done and the next time the user has to transmit it is
again done in the same manner. In the second case, the priority users compete amongst each other
for channel access, they do this by jamming the channel using the BBs, proportional to the waiting
time. Which effectively means that the user that has been waiting the longest obtain the channel. The
two events also have a priority between them; if there are scheduled priority users they will transmit
before the users that are of type (i). This is ensured by making the time before the black-bursts are
transmitted less than the time the users of type (i) sense if the channel is empty before going into
contention by means of the Contention Window. This means that the BB mechanism can be applied
on top of a normal CSMA/CA algorithm. The Black Bust mechanism is depicted in Figure 3 .9.
U1 U2 Data U1 U2
Channel
User 1
User 2
= Black Burst = Data = Scheduled

transmission transmission transmission
Figure 3.9 Basic operation of the Black Burst Scheme.
Page 25 of 65
IST-2001-34157
Although there is some scheduling of the priority class users, the mechanism is still random. The
disadvantage is that the priority users have ‘hard’ priority over the best-effort users [15]. The work is
however done with some relatively strong assumptions. The most important one being that the GI and
GS are identical during the analysis, meaning that all interfering nodes can be sensed, thus effectively
assuming that the analysis is done in absence of the hidden terminal effect. If the BB system is put in
a situation where hidden nodes do exist, the performance is severely degraded by their presence.
One important issue for investigation is the interaction between the MAC layer and the higher layer. In
order to investigate this in [15] two different operation modes are defined:
 Operation with feedback: A RT source can generate blocks of data at regular intervals. The
MAC layer can convey all information blocks created up to a certain moment in a packet and
transmit the packet when the channel access instance comes. This way the packet size varies
with the encountered delay. In practice however it is difficult to create packets in this manner.
 Operation without feedback: this is a simpler communication architecture, where already
assembled packets are passed onto the MAC layer for transmission one-by-one.
This is also the case when a node acts as a relay station. Here the operation without feedback is
assumed, since this is the most general one (both originating and relaying sources can apply it), and
the most feasible one from an implementation point of view.
Considering the case where a RT-node has an inter-arrival time for the packets to be transmitted of
trdy. To keep the delay bounded, the time between two scheduled transmissions of packets ( tsch )
should be less than the inter-arrival time, especially since the channel may be occupied by other RT
sources or best-effort traffic at the scheduled instances. Now tsch can be found as:
tsch = trdy - tbslot - tobs - 
with  > 0.  is being called the slack time and should be larger than the time it takes to transmit a
maximum length data packet since on the scheduled instance it is possible that a BE transmission
prevents the other users from transmitting immediately.
3.1.2 MAC layer for H-OFDM
The H-OFDM system as it is being considered in the PACWOMAN project (WP4.2) uses a hybrid of
two multiple access schemes: Orthogonal Frequency Multiple Access (OFDMA) and Time Division
Multiple Access (TDMA). This means that the OFDM signal is frequency and time shared in order to
achieve a larger granularity of bit rates than is achievable with pure OFDMA or TDMA. This higher
granularity allows for a large flexibility and scalability on the high data rate links.
The parameters that characterise the H-OFDM system are thus of two natures. There are parameters
that describe the OFDM signal and parameters that describe the TDMA frames. In OFDM the total
available bandwidth is divided into N subcarriers, which are orthogonal. The symbols that are to be
transmitted are multiplexed in frequency by modulation on the subcarriers. To prevent adjacent
channel interference some of the subcarriers can remain unused as frequency guards and some can
be used for channel estimation.
The OFDM parameters may be adapted depending on the Channel State Information (CSI), the main
parameters than could be modified are:
 The number of subcarriers N
 The length of the OFDM symbol T
 The number of frequency guards Nfg
 The number of pilot sub-carriers Np
 Length of the cyclic prefix
Page 26 of 65
IST-2001-34157
Note that the number of available or useful carriers (Nu) is influenced by these choices: Nu = N – Nfg –
Np. Note however that the parameters N and T cannot be changed independently if the bandwidth and
cyclic prefix are given
The multiple access features for OFDM are introduced by assigning a subset of the N subcarriers to a
certain user. For example when one subcarrier is assigned to each user the system can accommodate
N users. In the H-OFDM proposal the number of sub-carriers that can be assigned to a user i is
flexible and referred to as si. The concept is depicted in Figure 3 .10.
B
s=5
Figure 3.10 Division of the N sub-carriers over multiple users with different data rates.
The bit rate r of a sub-carrier depends on the modulation alphabet, for now this is assumed to be
constant. However a flexible modulation alphabet would increase the degrees of freedom in the
system and thus the adaptability. The H-OFDM proposal will apply convolutional coding to exploit
channel diversity.
There is also a TDMA structure; the OFDM symbols are grouped into time slots, which have a
header/preamble. Frames consist of between 0 and l of these slots, with a maximum of L. This
structure is depicted in Figure 3 .11.
FR i-1 FR i FR2 i+1

Super-frame level
TS 1 TS 2 TS L
Frame level
Header
OFDM symbol OFDM symbol OFDM symbol
(preambles)
Slot (packet, burst) level
Figure 3.11 Illustration of slot, frame and super-frame levels.

Depending on how the time slots are grouped, different bit rates can be achieved. Also the polling
frequency influences this, how often is a frame assigned to a certain user.
With these two sets of parameter, i.e. the OFDMA and TDMA ones three different approaches for Link
Adaptation are proposed:
1. Assigning all sub-carriers to a single user during some slot (si = Nu), the data rate can be
varied by l.
2. Assigning some sub-carriers to a user for the whole duration of a frame (l = L), the data rate
can be varied by changing s.
Page 27 of 65
IST-2001-34157
3. Assigning some sub-carriers to some slots, the bit rate can be determined by varying both s
and l.
The first option is basically a combination of OFDM and TDMA. The second option is OFDMA with
TDMA and the third option is a full hybrid with complete flexibility.
3.2 Design Considerations
In general the requirements from Section 3.1 and D2.2 should be met. However the two considered
technologies have their own additional peculiarities.
3.2.1 Extensions to CSMA
The carrier sensing based protocols can be executed in a completely distributed manner. There is no
need for a centralised node that schedules the transmissions. This allows for a large flexibility in the
system, nodes can leave and enter at their convenience without the need to register and deregister.
The exponential backoff mechanism adjusts automatically the size of the contention window
depending on the size of the population of communicating terminals (more communicating terminals
correspond to more collisions).
However, this flexibility comes at a price, for larger number of nodes the MAC scheme comes in a
collision limited domain and the performance is degraded severely. Additionally the basic carrier
sensing MAC introduces unfairness, the last successful node has an advantage over the other nodes,
since its backoff window is smaller
Another performance degrading factor is are hidden terminals, which are not sensed by nodes that
want to communicate, causing collisions. Similarly exposed terminals prevent nodes from transmitting
when at the receiver no collision would occur [31].
Finally at the moment in IEEE 802.11 there is no possibility for association with two BSSes with one
interface, two network cards required for this. This means that it is quite inefficient to use this interface
for true ad-hoc networking.
3.2.2 H-OFDM
The H-OFDM scheme is more static due to the use of TDMA. A centralised entity is required and
complete distributed implementation of the MAC scheme is impossible. However the mechanism is
more efficient for higher data load.
Also many levels of synchronisation are required, first of all at the physical level since OFDM is very
sensitive to offsets in the synchronisation. But also at the MAC level synchronisation is required, users
need to know which blocks of subcarriers are assigned to which user. If they want to start to
communicate, also the length of the time frames should be determined, both for the own transmission
as of the other communicating users to be able determine when the channel becomes available again.
To apply carrier-sensing mechanisms all the sub carriers in the block assigned to the user should be
monitored. If such a block was not assigned then the whole spectrum can be monitored to see if there
are sufficient subcarriers available for the transmission. To prevent these sensing operations, some
logical link (subset of timeslots and subcarriers) could be dedicated as control channel to exchange
the parameters required for the transmission.
In [33] a proposal has been made for a leader election procedure based on the H-OFDM concept.
When following this procedure a terminal is elected as a leader and this terminal is used to
synchronise and co-ordinate the transmissions. For this purpose a Leader Channel (LCH) has been
defined, which is a logical channel mapped on the first slot of every frame.
One advanced feature of the MAC is the implementation of load dependent degradation: if a high rate
user wants to enter the system and the system is already quite loaded. The bandwidth assigned to the
Page 28 of 65
IST-2001-34157
newly entering user and possibly some of the other users is reduced so that everybody has at least
some level of service. Applicability of this feature depends on the user requirements to the connection.
3.3 MAC alternatives
In this Section several options for extending the CSMA protocol to make it better suited for a multi-hop
environment are considered. The extensions are based on channel jamming using energy pulses as is
for example done in the Black Burst extension of CSMA[15]. A receiver-initiated version of the
BlackBurst protocol for multi-hop scenarios is proposed.
Another system based on energy pulses has been proposed by Papachristou and Pavlidou in [16].
Collisions are avoided by transmitting short sequences of energy bursts that are transmitted as busy
tones. Each node produces a unique pattern of tones. Each node i counts the number of successful
transmissions ai, since its last successful transmission, modulo the total number of nodes N. Since no
two nodes can transmit successfully at the same time each node counts a unique number of
successful packets ai. This number is used as the unique binary identifier. The identifier is transmitted
by on-off keying the blackbursts (for each ”1” a energy pulse is transmitted, for a ”0” the station listens
to the channel) starting with the most significant bit. The user with the highest identifier (waiting the
longest) captures the channel. The disadvantage of this method is that the exact number of nodes
needs to be known to determine the optimal length of the contention period, further all nodes need to
1. Black Burst combined with RTS/CTS: The most straightforward modification is the integration
of BB with the RTS/CTS mechanism. However the additional protection offered by the BB
mechanism is limited in this case.
2. Black Burst combined with FAMA: A black burst energy pulse can be used to protect the RTS
within the FAMA scheme. The CTS is already protected by an additional length. This way the
collisions of RTS messages are reduced.
3. Coded Black Burst: On-off keying with BB-burst to transmit additional information, e.g. to
replace RTS.
4. Multicast Black Burst: each node that receives a one hop BB broadcasts it ones. Jamming,
both the transmitter and receiver area. Geographically very inefficient.
5. Setting the NAV based on Black Burst: Does not provide complete protection. Requires
modulation of the Black Burst pulses.
6. Making the mechanism receiver initiated: After the RT flow starts, it can be assumed that
packets arrive at regular intervals. The system can be made receiver initiated. Sending and
CTS to reserve the channel at the moments that a packet is expected. The advantage is that
we can use 'exposed' transmitters, so geographically more optimal solutions, still efficient
mechanism [35].
Therefore here we evaluated the last method and compared it with the basic operation in a single and
multi-hop environment [32]. The basic idea here is to move the black burst scheme on the receiver-
side and to combine it with a floor acquisition technique. The suggested solution solves the hidden-
node problem, enforces a round-robin discipline among real-time nodes and gives priority to real-time
traffic, thus achieving QoS requirements. The proposed access scheme will be referred in the
following as BlackBurst++, or simply BB++.
Using bursts of energy to resolve access contentions among priority nodes has proved to be an
excellent technique, with two main advantages:
1. Every contention has a unique winner

2. It's completely distributed
The fact that it is completely distributed is of the utmost importance in ad-hoc networks, where it is not
desirable to have a central coordinator to control traffic flows, so that the network can self-organize
and operate in every condition. The RTS/CTS technique is also the building block of many well-
studied solutions to the hidden-node problem. The BB++ scheme gets rid of the RTS control frame
Page 29 of 65
IST-2001-34157
and uses a CTS modified frame as an invitation for real-time nodes to send their data. This reduces
the signalling load and thus the overhead of the protocol.
In combining these two techniques we take advantage of the good properties of both of them. The
BlackBurst++ scheme makes an effort in this direction, there are still many design issues to address,
but preliminary simulations and an analysis of the proposed scheme suggest that this direction is
worth to be followed.
3.4 Implementation Proposal
3.4.1 Basic Protocol Operation
Real-time flows have an important characteristic: they are predictable. Assuming that real-time
packets are generated every trdy units of time, the receiver of a real-time flow can take advantage of
that. After receiving the first packet of a real-time flow, the receiver can assume that new packets will
arrive (approximately) every trdy units of time. On the other hand, the same receiver doesn't know if at
the time when a RT-packet is expected, it will be busy in a communication with another neighbour
sending, i.e., data traffic.
What the common receiver can do is to protect the arrival of a real-time packet with a Clear To Send
(CTS) control frame ensuring that none of the neighbouring nodes transmit simultaneously. The CTS
frame is an invitation to the real-time node to send its packet. At the same time, the CTS acquires the
floor for the incoming RT-packet, any other node in the range of the receiver, but (eventually) not in
the range of the real-time user is inhibited to transmit for a period of time specified in the CTS frame.
This time is equal to the maximum duration of a real-time packet transmission. Neighbour nodes that
receive a CTS frame, set their Network Allocation Vector according to that duration. Hence, the real-
time packet can be sent free of collisions.
The basic signalling of the BB++ scheme is depicted in Figure 3 .12. Node A initiates a
communication session with Node B by sending a RTSb, with a flag set that the BB++ protocol will be
applied. If this message is received correctly by Node B, then this Node replies with a CTSa. Now
Node A can send a data packet to Node B. From this moment onwards node A and B have
established a BB++ context and node B will schedule regular Invite-to-Send messages (ITS), that are
transmitted to Node A indicating that it can send the next data packet. An Invite-to-Send is composed
of a CTS packet protected by a Black Burst transmission. Nodes to which the ITS is not addressed
refrain from communicating (here Node E and C).
b.
c. d.
Figure 3.12 Basic Protocol Operation.
Page 30 of 65
IST-2001-34157
A hidden node D is out of range of the ITS message, however it will sense the channel busy when
Node A transmits its data and refrain from transmitting at those moments. If the ITS collides with an
ongoing transmission in range of Node A, e.g. a data transmission or a RTS. Then the invite is resent
later, since the receiver can see from an expiring time out indicating the no data is arriving as a
response to the ITS that the message was lost. Node A has priority over all non-RT nodes since a
shorter time-out is applied, and with the RT nodes the BB++ scheme divides the bandwidth fairly. So
as in the original BB and additional delay of one BE packet has to be taken into account.
Referring to the typical hidden-node scenario shown in Figure 3 .12d, how the BB++ access scheme
works in this case is explained in detail by the time diagram in Figure 3 .13.
Title:
f2pspost.dvi
Creator:
dvips(k) 5.90a Copyright 2002 Radical Eye Software
Preview:
This EPS picture was not saved
with a preview included in it.
Comment:
This EPS picture will print to a
PostScript printer, but not to
other types of printers.
Figure 3.13 The proposed scheme - blackburst++.

Both the transmitted (Tx) and received (Rx) packets at a given node are shown on the same axis (y >
0 for Tx packets, y < 0 for Rx packets). Node 1 and node 2 are hidden from each other and they both
need to communicate with node 0 in the middle, so they both receives node 0's transmissions. Node 2
has bursty data traffic to send, while node 1 has real-time, delay-constrained packets to send every trdy
seconds. Node 1 sends its first real-time packet using the usual DATA/ACK exchange with the
destination. After receiving the first packet of a real-time flow, the destination assumes that a new
packet will arrive in the near future (approx. in trdy seconds) and schedules an invitation for node 1 in
tsched seconds from the instant a real-time packet has been correctly received. The invitation (ITS,
composed of a BB+CTS should arrive as soon as the sender has a new packet available for
transmission. On the other hand, the sender will wait until it is invited to transmit by node 0, and defers
without trying to access the channel until the ITS arrives. On the scheduled invitation attempt, if the
channel has been idle for tmed consecutive seconds, node 0 can successfully send its ITS frame, just in
time for node 1 to transmit the newly arrived real-time packet. Node 2 is hidden to node 1, but it
receives the BB+CTS frame sent by node 0 and can allocate its NAV accordingly 2, thus avoiding to
collide with the incoming real-time packet at node 0. Moreover, the inter-frame space between a BB
and a CTS is always tobs, which guarantees that the CTS is not going to collide with any other data
packet from a neighbour in the radio range of node 0. Hence, despite it is not its main usage, the BB
also works as a protection for the CTS. The real-time packet is sent free of collisions, and the hidden-
node problem is alleviated.
It can happen, that on a scheduled invitation attempt, the air around node 0 is busy with a data packet
transmission from a node hidden to the real-time user. Since, the latter will not start a packet
transmission unless invited, there is no chance that its packet collides with the currently received data
packet at node 0. When the data packet reception is completed (DATA+ACK), node 0 will wait until
the channel is idle for tmed and then send its BB+CTS. A new data packet transmission cannot start in
2
For now, it is assumed that the channel doesn't introduce any error, so that every packet is correctly
received if doesn't collide with another packet. With this assumption, node 2 always correctly receives
a CTS
Page 31 of 65
IST-2001-34157
the middle, because data nodes wait for a longer inter frame space (DIFS) before re-attempting to
grab the channel. This is how the BB++ RT stations have priority over best-effort. Thus, as soon as
the data packet frees the channel, the real-time user has priority access to the medium. Now the
length of the black burst is longer because the receiver has experienced a longer delay while
attempting to acquire the floor.
In this case the access delay ( w ) experienced by the real-time user(s) can not necessarily be
neglected, however it can be shown that it is bounded. In fact, in the worst case, the real-time packet
is available for transmission at the sender at the same instant a data packet transmission starts. So, at
maximum, it has to wait until a maximum length data packet transmission time, plus some fixed
timings:
w  tdata + SIFS + tACK + tmed + bmax + tobs + tCTS + SIFS
where bmax is the maximum burst length associated to a contention delay dmax = tdata + SIFS + tACK + tmed.
Obviously, the choice of tsched is a critical design issue for a proper behaviour of the scheme. Beside it
is still an open issue, some design evaluations on this point will be discussed later.
3.4.2 Contention resolution
Consider now the situation shown in Figure 3 .14, where two real-time receivers may need to
synchronize themselves in such a way that they will not disturb each other's reception of a real-time
packet. In Figure 3 .14, Rx1 and Rx2 are both destinations for two separate real-time flows. The two
flows starts from Tx1 and Tx2, which are hidden to each other. By using the Black Burst contention
mechanism, Rx1 and Rx2 can organize themselves in a TDMA structure, with the one experiencing the
longest contention delay being sure to win an eventual contention with other receivers.
Title:
f2pspost.dvi
Creator:
Preview:
Comment:
Figure 3.14 Contention resolving - A sample scenario.
In Figure 3 .15 it is shown how the contention resolution mechanism is employed. On the left side of
the diagram, the receiver-oriented scheme described above is enforced either by Rx 1 and Rx2, until
(on the right side) they are not disturbed by a data packet transmission, that may arrive from a fifth
node not shown in the figure. Both receivers have an invitation attempt scheduled for that time but,
since the channel is busy, they defer until the channel can be detected idle for tmed consecutive
seconds. At this later time, they both start a contention phase by jamming the channel with a number
of black slots proportional to the experienced contention delay. It is exactly as the original Black Burst
scheme, but receiver oriented. The length of the burst is still given by the following equation:
Page 32 of 65
IST-2001-34157
 d 
lbb  1   cont   tbslot
  tunit  
After sending their bursts, they both observe the channel for tobs seconds, to see if any other contender
is sending a longer burst: Rx 1 will win the contention because it has experienced a longer delay ( dcont1
> dcont2 ), thereby transmits a longer burst, and can transmit its invitation.
Title:
f2pspost.dvi
Creator:
Preview:
Comment:
Figure 3.15 The proposed scheme - Contention resolution.
The contention phase will lead to a unique winner. In fact, as in the original version of the protocol,
also in this BB++ scheme two scheduled invitation attempts are staggered in time by at least tpkt, the
transmission time of a maximum length real-time packet transmission, and so are their contention
delays. Therefore, counting the burst length in terms of tpkt (i.e. choosing tunit = tpkt ) ensures that two
different receivers always contend by mean of black burst that differs for at least one black slot.
The BB++ scheme described above, purposes a solution to the hidden node problem that is also able
to give priority access to real-time traffic. The main advantages of this scheme are, that real-time
packets are sent free of collisions and that they are no longer affected by the presence of hidden-
nodes. Further,
 Multiple real-time flows are organized in a TDMA structure without requiring a central
coordination (totally distributed).
 Access delay for real-time flow is bounded.
 High real-time throughput because the scheme is ideally collision-free.
3.4.3 The choice of tsched
The main critical design parameter in the proposed scheme is finding an appropriate value for tsched. It
is a critical choice, because from this parameter depends the correct behaviour of the protocol and its
stability. Clearly, tsched should be chosen such that the access delay of real-time packets always stays
bounded. Referring to Figure 3 .16 an analysis of the scheduling time is made.
As can be seen from Figure 3 .16 the scheduling time ( tsch ) is defined as the time between the end of
a real-time transmission and beginning of the contention for the next transmission.
Page 33 of 65
IST-2001-34157
Title:
f2pspost.dvi
Creator:
Preview:
Comment:
Figure 3.16 The proposed scheme - The choice of tsched.
The source node (SRC) generates new packets every trdy units
of time, with nth packet being generated at time:
t(n) = t(1) + (n - 1)  trdy (3.1)
The access delay w(1) of the first packet, that cannot be

predicted, is given by:
w(1) = u(1) - t(1) (3.2)
a first goal in the design of tsched would be choosing it in order to recover from the initial delay, reducing
the access delay of the 2nd packet, and leading to the smallest possible access delay for subsequent
packets. The condition is:
w(2) < w(1) (3.3)
where w(2) = u(2) - t(2), and the access instant of the 2nd packet is given by:
u(2) = s(2) + ( tbslot + tobs + tCTS + SIFS )

= u(1) + tpkt + tsched +  (3.4)
where,  = ( tbslot + tobs + tCTS + SIFS) is a fixed quantity of approximately 500 to 600 s. Substituting the
above equations (3.1), (3.2) and (3.4) into (3.3), an upper bound can be obtained for tsched:
tsched < trdy - tpkt - 
So, a possible choice would be introducing a parameter to satisfy this bound. For coherence with the
standard black burst scheme, this parameter will be called the slack time  > 0 and then tsched is given
by:
tsched = trdy - tpkt -  - 
With this choice for tsched it can be easily shown that starting from the second packet the access delay
decreases from the initial value with a rate that depends on . The slack time should also be upper-
bounded so that tsched > 0 at every time.
Page 34 of 65
IST-2001-34157
4 MEDIUM ACCESS CONTROL FOR THE LOW DATA RATE LINK
This section is dedicated to the medium access control layer for the low data rate link (LDR) of the
PACWOMAN system. Within the PACWOMAN project, the suitability of UWB technology for this low
data rate link is evaluated. UWB-based systems are most likely to operate in the frequency band
between 3.1 and 10.6 GHz (see also [7]) with a nominal bandwidth of at least 500 MHz. With respect
to the MAC layer it is important to note that it is very likely that UWB systems are going to operate in
time division duplex (TDD) mode. Furthermore, given the relatively large bandwidth and the low power
spectral density (the upper limit is set to –41.25 dBm/MHz, see [7]), UWB systems are vulnerable to
interference from narrowband systems operating in the same frequency range as the UWB system.
Therefore, interference detection and avoidance mechanisms, which are not an issue for conventional
wireless systems, are important aspects in the context of the MAC layer design for UWB-based
systems.
In this deliverable we will briefly review the general requirements for the low data rate link and based
thereon the more specific constraints for the MAC layer are derived. Furthermore, we are going to
detail the different design aspects and discuss the suitability of MAC layers of existing and emerging
wireless communication systems. The actual design and performance analysis of the MAC layer for
the UWB-based LDR component will be carried out in the PACWOMAN deliverable D4.3.1.
4.1 Requirements and Constraints

The LDR system is envisioned to operate in the so-called body-to-body as well as in the body-to-
access point environment (see also D2.1 for a more detailed description of the system requirements).
The body-to-body environment shows a maximum link distance in the range from one to two meters
whereas the maximum link distance of the body-to-access point environment ranges from three to five
meters. The system is operating in the regulated UWB frequency bands ranging from 3.1 to 10.6 GHz.
Accordingly, the system is expected to experience interference from narrowband systems also
operating in these frequency region, like e.g. the WLAN systems operating at 5 GHz.
Typical applications for the LDR system include the transmission of data from body-worn sensors to a
central and also body-worn terminal (body-to-body environment). Depending on the sensors, the
amount of data to be transmitted varies from several tens of bits (e.g. for a temperature sensor) to
serval kbits (e.g. for electrocardiogram data). Two basic types of services are distinguished, namely
the very low rate service and the low rate service. The associated data and tolerated bit errror rates
range from 100 to 10 kbps and 10 -3 to 10-7, respectively. The tolerated delay is in the order of one
second, thus for both services. Note that the support of voice traffic is optional. Furthermore, it is worth
noting that data exchange takes place in a rather sporadic and infrequent fashion, thus resulting in a
relatively low overall traffic.
The basic mode of operation is in master-slave mode, where the central body-worn so-called smart
terminal coordinates the data transmissions. An optional ad-hoc communication mode is also foreseen
for communication among multiple smart terminals. The amount of maximum users that can be
supported simultaneously is set to ten.
A low overall power consumption is of foremost importance and it is anticipated that terminals
operating in body-to-body environments run on a single alkaline battery for at least one to two years.
This constraint especially applies to the sensors terminals. The smart terminals, like e.g. the central
master terminal, are expected to run or more sophisticated and expensive batteries that can be
recharged or exchanged. Therefore, terminal capable of operating in the body-to-access point
environment are granted a power autonomy of one to two months.
The system requirements relevant for the MAC layer design of the LDR component of the
PACWOMAN system are summarized in the following table.
Page 35 of 65
IST-2001-34157
very low rate service low rate service

Amount of data 100 bits 10 kbits
Data rate 100 bps 10 kbps
Required BER 10-3 10-7
Tolerated delay 1s
Number of users maximum of 10 simultaneous users
Body-to-body environment: 1-2 years
Autonomy
Body-to-access point environment: 1-2 months
Table 4.2: MAC related system requirements for envisioned LDR services.
4.2 Design Considerations
In general, medium access protocols for wireless communication systems are designed with respect
to network performance metrics such as throughput, efficiency, quality-of-service support, and packet
delay. Given the above-presented constraints, energy consumption at the MAC level is an additional
major performance metric that has to be taken in consideration during the design process for the
development of the LDR component of the PACWOMAN system. The premise is that dumb terminals
have a very limited power, whereas the smart terminals are provided with larger power reserves.
The chief sources of energy consumption due to MAC related activities are the CPU, the transmitter,
and the receiver. CPU usage in the dumb terminals may be reduced by relegating most of the high-
complexity computation related to medium access to the smart terminals. Therefore, the focus of the
medium access protocol design and optimisation is on transceiver usage. The radio can operate in
three modes: standby, receive, and transmit. In general, the radio consumes more power in the
transmit mode than in the receive mode, and consumes least power in the standby mode. In general,
four directions to reduce the energy consumption have been identified [8]:
1. Collisions should be eliminated as far as possible since they result in retransmissions leading
to unnecessary energy consumption and also to possible higher delay. Note, however, that
retransmissions cannot be completely avoided in cases of high link error rates. This aspect is
also of importance in collision-free medium access schemes, where new terminals registering
with the master terminal may have to use some form of random access protocol. Techniques
such as reservation and polling can help meet the requirement that collisions be minimized.
2. Protocol overhead should be kept as small as possible. The size of the control packets has to
be adapted to the size of the data packets in order to yield a sufficient efficiency. The time
associated with protocol overhead depends on the synchronisation and channel estimation
requirements of the PHY layer, the synchronisation requirements of the MAC layer, as well as
on the amount of necessary control information for the protocol itself. On the other hand, the
time spent to transmit the actual data depends on the spectral efficiency of the PHY layer and
the amount of data to be transmitted, which in turn is application-dependent. Especially for low
data rate applications, where small amounts of data are exchanged, the reduction of the
protocol overhead is of foremost importance and subject of a meticulous design.
3. The time spent in receive mode should be reduced as much as possible. In common wireless
communication systems, the receiver has to be powered on at all times, e.g. to analyse the
data received control information (to determine whether the information is for other terminals
or not) or to keep track of channel status through constant monitoring. This results in
significant energy consumption. Two different categories can be identified:
a. Listening to a data transmission that is destined to someone else, also referred to as
overhearing. Techniques to prevent overhearing have been presented in [8] and [9].
b. Listening to the transmission channel during time intervals without data transmission,
which is also referred to as idle listening. Especially to reach terminal lifetimes of
several months or even years with a single alkaline battery, the transceiver will have
to remain in standby mode most of the time.
Page 36 of 65
IST-2001-34157
In the context of the envisioned LDR component of the PACWOMAN system all the above-mentioned
points will have to be considered in order to reach the requirements. However, since efficient
techniques reducing the number of collisions (e.g. collision avoidance) and preventing overhearing are
known, special emphasis will have to be given to the second and the last point (3b). Especially the
small amounts of data as well as the rather low traffic load will have a crucial impact on the
performance of the medium access control. Concerning point 3b, it is anticipated that the use of
TDMA-based protocols provides an efficient means to avoid idle listening [10].
Furthermore, it is anticipated that the aspect of interference detection and appropriate interference
avoidance or mitigation strategies play a major role for the design of an adequate MAC layer for the
UWB-based LDR component.
4.3 MAC Alternatives
The MAC layers of WLAN systems like e.g. the IEEE 802.11 standard are not suitable for the LDR
component of the PACWOMAN system since the primary optimization criterion of these MAC layers
was the to maximize the throughput and to minimize the delay. A low power consumption was not
considered as a primary design goal. This is reflected by the fact that WLAN systems need to listen
most of the time in order to determine whether the channel is free or not, which in turn amounts for a
huge portion of the overall power consumption.
As outlined in the previous section, reduction of the protocol overhead and idle listening are of
foremost importance for the MAC layer design of the LDR system component. It is anticipated that the
IEEE 802.15.4 could effectively limit idle listening due to its TDMA component. However, due to the
differences in the PHY layers the protocol is not directly applicable. Especially the PHY specific
synchronisation procedures and the associated protocol issues (preamble design) need to be
evaluated carefully. Furthermore, the use of the IEEE 802.15.4 protocol needs to be investigated in
the context of the envisioned LDR scenarios in order to derive an adequate parameter set that
optimizes the overall power consumption. These and other issues, like for instance the use of simple
ALOHA protocols for the ad-hoc option, are going to be studied and evaluated in the PACWOMAN
deliverable D4.3.1.
An initial study has been performed to consider more flexible MAC layers that could be used in an ad-
hoc manner. The MAC layers apply link adaptation to adjust themselves to the interference conditions
of that moment and can adjust the setting to a varying number of nodes. In a master-slave setting with
a limited number of nodes these advanced features might not be required but in future scenarios with
larger numbers of nodes they are mandatory. The results of this study can be found in Appendix A
[22]. Where a time-hopping spread spectrum (TH-SS) UWB communication system is considered and
it is discussed how to design medium access techniques in order to achieve efficient resource sharing
in WPAN environments. The throughput performance is evaluated by means of computer simulation in
a multi-network scenario, and the impact of TH parameters (such as the number of bins per frame and
the processing gain) on the throughput performance is discussed. Furthermore, a link adaptation
mechanism is proposed which can adapt TH parameters according to interference conditions among
collocated WPANs. This mechanism aims to maximize the throughput achieved by WPANs while
keeping the robustness with respect to the basic timing mechanisms used in the TH-SS
communications. The proposed mechanism is based on chip puncturing, i.e. transmitting less chips
than determined by the hopping code, leaving some of the chips unused and thus reducing the
interference in the system. The work done so far is an initial exploration of the topic, identification of
the most important parameters and development of the link adaptation mechanism.
Page 37 of 65
IST-2001-34157
5 SIMULATION TECHNIQUES
5.1 Review of possible simulation platforms
5.1.1 OpNet
OpNet was originally developed at MIT and introduced in 1987 as the first commercial network
simulator. OpNet is a comprehensive software environment for modelling, simulation, and analysing
the performance of communication networks. OpNet modeller is considered to be one of industry’s
leading edge network simulators, which allows the user to design and study network equipment,
communication protocols, and systems.
OpNet has demonstrated a richness of features. The built-in functions assist the user in the model
specifications through a very well designed user-interface. This user-interface is based on graphical
editors, which provide an intuitive mapping from the modelled system to the OpNet model
specification. The pre-defined functions are very useful to control precisely the event management.
The object orientation supports the definition of new classes and objects. The re-usability of those
objects is emphasized by their set of attributes. OpNet also provides additional tools to analyse and
graphically represent the simulation results.
On the other hand, the main disadvantage is its relative complexity to model a given system. The time
required to achieve the modelling of a system can be very long, especially for new developments.
Furthermore, it is reported that the OpNet simulator is pretty memory consuming and that it is difficult
to modify the library models. Finally, since OpNet is a commercial tool, purchase of the software as
well as the model libraries are expensive.
Supported platforms are Windows 2000, Windows NT, and Unix. Additional information can be found
on the OpNet site http://www.opnet.com/ .
5.1.2 Network Simulator 2
NS/2 is the second version of the network simulator (NS). NS is a discrete event simulator targeted at
networking research. NS provides substantial support for simulation of TCP, routing, and multicast
protocols over wired and wireless networks. NS began as a variant of the REAL network simulator in
1989 and has evolved substantially over the past few years. In 1995 NS development was supported
by DARPA through the VINT project at LBL, Xerox PARC, UCB, and USC/ISI. Currently NS
development is support through DARPA with SAMAN and through NSF with CONSER, both in
collaboration with other researchers including ACIRI. NS has always included substantial contributions
from other researchers, including wireless code from the UCB Daedelus and CMU Monarch projects
and Sun Microsystems.
The NS distribution includes a number of validated protocols, such as different MAC layers for local
area networks (LAN). The MAC protocols currently supported include CSMA/CA (modelled after the
AT&T WaveLAN), CSMA/CD (modelled after standard 10 Mbps Ethernets), Multihop (loosely
modelled after Metricom's Ricochet), and the MAC protocol of the IEEE standard 802.11. Finally, there
are a number of contributed protocols described, which can be found on the site
http://www.isi.edu/nsnam/ns/ns-contributed.html. Note, however, that these protocols are often for
specific releases of NS and may not work in the current release.
NS is freely available and has been developed for several kinds of Unix (FreeBSD, SunOS, Solaris),
Linux, and also Windows. Simple scenarios should run on any reasonable machine, but very large
Page 38 of 65
IST-2001-34157
scenarios benefit from large amounts of memory. Additional information can be found on the NS/2 site
http://www.isi.edu/nsnam/ns/ .
5.1.3 GloMoSim
GloMoSim is a scalable simulation environment for wireless and wired network systems. It is being
designed using the parallel discrete-event simulation capability provided by Parsec. GloMoSim
currently supports protocols for a purely wireless network. In the future, it is anticipated to add
functionality to simulate a wired as well as a hybrid network with both wired and wireless capabilities.
Most network systems are currently built using a layered approach that is similar to the OSI seven
layer network architecture. The plan is to build GloMoSim using a similar layered approach. Standard
APIs will be used between the different simulation layers. This allows the rapid integration of models
developed at different layers by different people. The protocols being shipped with the current library
support the following MAC layer mechanisms: CSMA, IEEE 802.11 and MACA.
The GloMoSim environment relies on the Parsec compiler, which is included in the GloMoSim
distribution. For the development of custom protocols in GloMoSim, some familiarity with Parsec is
required. Most protocol developers will be writing purely C code with some Parsec functions for time
management. Parsec code is used extensively in the GloMoSim kernel, but it is not required to know
and understand how the kernel works.
Two versions of the simulation tool exist: the academic research version, which is for academic use
only, and a commercial version, which is distributed as the QualNet software package. The latter is
available for Windows 2000 and XP, Linux and Unix (Solaris with SPARC architecture). Additional
information can be found on the GloMoSim site http://pcl.cs.ucla.edu/projects/glomosim/ as well as on
the Parsec site http://pcl.cs.ucla.edu/projects/parsec/.
5.1.4 Ptolemy II
Ptolemy II is a set of Java packages supporting heterogeneous, concurrent modelling and design. Its
kernel package supports clustered hierarchical graphs, which are collections of entities and relations
between those entities. Its actor package extends the kernel so that entities have functionality and can
communicate via the relations. Its domains extend the actor package by imposing models of
computation on the interaction between entities. Examples of models of computation include discrete-
event systems, dataflow, process networks, synchronous/reactive systems, and communicating
sequential processes. Ptolemy II includes a number of support packages, such as graph, providing
graph-theoretic manipulations, math, providing matrix and vector math and signal processing functions,
plot, providing visual display of data, data, providing a type system, data encapsulation and an
expression parser, etc. The simulator provides different simulation domains to model different types of
systems. For signal processing (Synchronous, Dynamic and Boolean Dataflow), telecommunication
switching software (Message Queue), communication network modelling (Discrete Event,
Communicating Processes) and control (Finite State Machines).
The website of the Ptolemy II project can be found online: http://ptolemy.eecs.berkeley.edu/ptolemyII/.

Ptolemy is being developed by UC Berkeley
5.1.5 Tailor-Made Tools
Input: short description of simulator and simulation principles, availability of software and libraries.
Discussion of advantages and drawbacks! Web-link: ...
WIPSim, C++ programming, Sanas
Page 39 of 65
IST-2001-34157
5.1.6 Discussion and Simulator Choice
This Section discusses the of advantages and drawbacks of the above presented simulators in the
context of the PACWOMAN objectives. In Table 5 .3 the presented simulation platforms are
compared at a high level.
Table 5.3 Comparison of different simulation platforms (H – High; M – Medium; L – Low).
Simulation tool Main parameters
Functionality Flexibility Scalability Complexity Exchangability
OpNet H M M M M
NS/2 H M H M H
GloMoSim M M H M M
Ptolemy II M L M H L
Tailor-made tools variable H H L L
The wireless networking support for Ptolemy II was found to be immature for use in the project and
does not have sufficient MAC functionality already implemented. GloMoSim has (for the moment) a
strong focuss only on wireless networks. The most elaborate tools are Opnet and NS/2. Of which the
latter is available freely on the Internet. It was decided, based on the available functionality, expertise
of the partners and it’s wide acceptance in the scientific world to use Network Simulator 2 for the
system level evaluations of the concepts.
Custom-made tools will used to test small algorithms and ideas, before a system level implementation
is done. This is because such programs have a very high flexibility and low complexity. Making it much
faster to implement and test small ideas and easier to study their effects. Because the exact
assumptions and simplification used in the tool are known.
5.2 Simulation Methodology and Scenarios
Topology refers to the layout of connected devices in a network. Many standard topologies exist,
including bus, ring, star, tree, and mesh. In order to simulate the Medium Access Control mechanisms
several different topologies can be used to investigate different aspects of the performance. Here we
will present them as simulation scenarios. Firstly we consider the single hop case, which is either a
point-to-point or star configuration. It is possible to include some hidden nodes in the analysis. A line
structure is a more generalised form of this, keeping the interference regular. Finally there is the
meshed graph with all nodes connected and the random graph with all nodes connected as well, but
with irregular (random) interconnections.
5.2.1 Single hop
The most simple simulation scenario is the one hop case; this basically
means that there is only communication with the direct neighbours (point-to-
point links). A typical evaluation scenario includes one base station and
several mobile terminals that communicate with the base station. All
terminals are within range of each other, so that a station can sense the
channel busy if another station is already transmitting when the station
wants to access the channel.
To keep the scenario simple, only a star topology is allowed. meaning that Figure 5.1: Single hop.
all users communicate with the same base station (sink). The scenario is
depicted in Figure 5.1, where a square is a base station and a circle is a
mobile station.
This scenario models multiple users that try to access the same link, while being able to sense all the
activity on the channel. It provides an upper limit for the performance, since it is not degraded by
activity on hidden links.
Page 40 of 65
IST-2001-34157
5.2.2 Single hop with hidden nodes
A more realistic scenario then the pure single hop case is the single hop scenario with hidden nodes.
Again within the main cell under investigation a star topology is used
where the mobile stations communicate with a base station.
However, it is no longer assumed that all stations are in range of
each other. As can be seen in Figure 5.2, there are nodes that are in
range of some of the communicating nodes, but not to all of them.
These nodes are call hidden terminals. Activity of these hidden
nodes cannot be sensed by all nodes in involved in the
communication, causing collisions on the link. These collisions
deteriorate the performance. Even in a single hop hidden nodes can
occur naturally, when two nodes communicate with the same base
station, but they are the maximum distance apart (twice the cell
radius). However, in a multi-hop environment it occurs even more
frequently because neighboring communications interfere with each
other. Figure 5.2: Single hop with
hidden nodes.
The scenario with single hop and hidden nodes is a first
approximation for the performance in the case of a multi-hop environment.
5.2.3 Line structure
The most simple and regular multi-hop connection consists

of a chain of inter connected terminals, the so called line
structure. Each node is interconnected with one neighbor on
either side (except the starting and ending node, this could
be avoided by using a ring structure, but then the
interference pattern potentially becomes more complex). In
this way it is quite easy to trace which nodes interfere with
each other. Because not all nodes are in range of each other
routing functionality is required to deliver the packets to the
destination. Using the line structure it is also possible to
Figure 5.3: Line structure.
investigate the interference effects of activity on links that are
not directly adjacent to the link under investigation. This is in
particular important when investigating handshake protocols
like RTS/CTS, since they can block transmissions on not adjacent links.
The line structure is close to single cell case; the interference is regular and orderly. It is a first step
towards the modeling of a realistic multi-hop environment.
5.2.4 Meshed structure
Mesh topologies involve the concept of routes. Unlike each of the

previous topologies, messages sent on a mesh network can take any
of several possible paths from source to destination. This means that
full routing functionality is required for this scenario and that there are
many possible interactions between the different links. However in a
real wireless ad-hoc system these interactions exist.
Figure 5.4: Full meshed
The meshed structure can be used to investigate the effect of many
structure.
simultaneous transmissions and how they interfere in a complex
environment and also to study the interaction between the routing
layer and the MAC layer and possible cross layer optimisations between them. For example to
investigate the effect of the flooding routing signalling over a particular MAC layer. The meshed graph
is made in a deterministic manner to investigate a particular scenario. It is also possible to create not
Page 41 of 65
IST-2001-34157
fully meshed graphs to model the case where no physical link exists between two nodes, for example
due to shadowing. A last method would be the generation of random graphs, not modelling a particular
scenario. But rather a class of connection types, with characteristics like scarce connections,
concentrated nodes etc.
5.2.5 Simulation Parameters
All the metrics considered are obtained averaging the performances of the wireless nodes belonging
to a certain group. The metrics have been chosen in order to evaluate if a given MAC protocol is able
to guarantee a good Quality-of-Service in terms of short packet delays, low percentage of packet loss
and efficiency. The main attention was given to evaluate the MAC layer performances, because a
crucial role is played by the access scheme in supporting delay-constrained services.
1. Throughput: The total throughput of stations enforcing a certain MAC protocol was
investigated. This metric gives a measure of the utilization of the wireless medium that is a
scarce resource, so an efficient use is very important. The aggregate throughput is the total
amount of data, expressed in Kb/s that is delivered to the destination. To compare graphs
from different levels of load a normalized throughput is introduced. The normalized throughput
is calculated as the fraction of the offered data that is actually delivered to the destination at
the application level (AGT).
2. Access delay: The access delay measures how fast the MAC layer does its job. It is
calculated as the time interval from when a packet is available at the MAC layer for delivery
and when it is actually delivered by the MAC layer on the channel. This interval is clearly
influenced by traffic conditions on the network and by the access scheme itself. The average
value of the access delay was considered. But sometimes, having an small {\it average} delay
is not enough for certain applications. There could be a maximum tolerable delay after which
the data is useless. A typical delay bound for voice traffic is 20ms, so the percentage of
packets exceeding this delay bound was also investigated.
3. MAC Efficiency: Knowing the throughput of a system gives an overview of how much of the
offered load was successfully delivered to the destination. A normalized throughput of one
means that 100% of packets transmitted by an application reached its counterpart on the
other hand of the link. The MAC layer is responsible to carry these packets on the channel
from source to destination, possibly in a fast and reliable way. Unfortunately, when many
wireless stations tries to access the medium at the same time, packets can collide with each
other or being dropped by the MAC layer due to errors introduced by the radio channel. Some
MAC layers, like CSMA/CA, recover from collisions by explicitly ACKing successfully received
packets and requiring a link level retransmission in case of failure. Other MACs leave this duty
to the transport layer. The more packets are lost under increasing load conditions, less
efficient the MAC layer is. The MAC Efficiency is calculated as the ratio between successfully
received packets by the destination's MAC layer and the total number of packets transmitted
from source's MAC. A value of one means that no packets are dropped on their way through
the destination and the MAC can efficiently deliver the entire offered load.
4. Jitter: One peculiarity of real-time traffic is that it is a continuous stream. Packets are sent at
regular intervals. The receiver expects to receive a new packet every trdy units of time. The
deviation from this ideal condition is measured by the jitter. In this specific case it was
measured as the difference between the expected packet arrival time and the actual arrival
time at the destination's MAC layer. A best-effort MAC mechanism like CSMA/CA gives no
guarantees those new packets will be delivered on-time, while QoS oriented schemes (like
Black Burst) do an extra effort to achieve zero or very low jitter.
Page 42 of 65
IST-2001-34157
6 MAC LAYER FORMAL DESCRIPTION AND MODELING
6.1 Description of Approach
6.1.1 Objectives
The approach described in the sequel is going to be used for the formal specification and functional
verification of the PACWOMAN MAC layer optimisation algorithms. SDL (Specification Description
Language) is going to be used for this purpose. More specifically, the objectives of the concerned
technical activities will be:
 The capturing of requirements as provided by the simulation-driven (NS/2, OpNet, etc) high-level
definition of MAC optimisations.
 The formal specification and placement of these optimisations in the context of existing MAC-layer
protocol formal descriptions (IEEE 802 family of protocols).
 The functional verification and optimisation of the produced SDL models.
 An evaluation of the implementation complexity induced by the proposed MAC-layer functions.
6.1.2 SDL for MAC-layer Formal Description and Early Verification
6.1.2.1 SDL Benefits
A system engineering process is commonly divided into few separate design phases between
requirement analysis and implementation. Until recently, the main programming tools for MAC
systems have been assembler languages. The increasing complexity has demanded more
sophisticated software design tools, such as optimised compiler, debuggers, and development
environments with simulators and verification tools. A solution to accelerate the design flow is to utilise
higher abstraction and formal languages. This facilitates the comprehending of complex systems and
enables hardware/software co-design.
SDL can be considered as a high-abstraction level programming language with a graphical user
interface. The language is especially beneficial in specifying real-time systems and generally
subsystems with a discrete stimuli-response type of behaviour, such as communication protocols. SDL
eases the verification of functionality, allowing e.g. to achieve high reliability with little target testing.
Moreover, it is possible to verify the system functionality at early states of the design.
The use of SDL facilitates a continuous design flow through the whole system development. By
extending and adding more details into a system SDL model, the output model of a previous phase
can be directly used as input for a next one. At the same time, the maintenance and redesign of the
system is simplified as the formal model can be used as documentation. Logic analysis, and
performance estimations can be applied in early phases for verifying the design.
6.1.2.2 SDL Principles
SDL is standardised by ITU-T according to the following requirements:

 Provide a well-defined set of concepts to be used for constructing a model for the specified
systems.
Page 43 of 65
IST-2001-34157
 Provide an unambiguous, clear precise, and concise specification, which is especially important
for avoiding mistakes and misunderstandings between separate bodies contributing to the system
specification.
 Enable the analysis of the specification for completeness and correctness.
 Provide possibility to utilise computer-based tools in order to create, maintain, analyse, and
simulate specifications.
SDL should also abstract the specification from the implementation in order to:
 Provide the basis for determining the conformance of implementation to specification.
 Give a clear overview of a complex system.
 Enable the postponing of implementation decisions, while not excluding any valid implementation
techniques.
The system behaviour in SDL is based on communicating extended finite state machines that are
executed concurrently. State machines are represented by SDL processes. Processes communicate
with each other and with the system environment by exchanging asynchronous signals that may carry
any number of parameters. Signals are usually trigger-type, although also continuous signals can be
used. SDL also contains timers that can be configured to generate signals at defined moments. Each
process in an SDL system contains a FIFO (First-In-First-Out) input buffer into which the received
signals are queued. When a process is in a state, i.e. not currently performing a transition, it is able to
receive a signal from the queue. Signal reception triggers a state transition.
SDL architecture contains a hierarchy that allows the dividing of large systems into more
comprehensible structures. The highest level, system level, consists of blocks and channels. Channels
connecting separate blocks and system environment, as well as routes between processes, are used
as signal carriers. Each block may contain any number of sub-blocks, with similar block and channel
architecture. The lowest level sub-block contains the actual processes. A process may have local
variables and procedures as attributes. Procedures are the lowest level in the functional hierarchy
containing their own local scope.
The dynamic behaviour in an SDL system is described in the combined behaviour of a number of
processes as shown in Figure 6 .17. A process is an extended Finite State Machine (FSM), which
operates autonomously and concurrently with other processes. The co-operation between the
processes is performed asynchronously by discrete messages, called signals. A process can also
send signals to and receive signals from the environment of the system. It is assumed that the
environment acts in an SDL-like fashion, and it must obey to constraints given by the system
description. The behaviour of a process is deterministic; it reacts to external stimuli (in the form of
signals) in accordance with its description.
Figure 6.17: System behaviour.
Page 44 of 65
IST-2001-34157
The aim of structuring is to cope with complexity. The process concept is also suitable for structuring.
In addition, a process must always be contained in a block, which is the main structuring concept. A
system contains at least one block. Channels interconnect the blocks. Moreover, the boundaries of the
system are also connected to the system's blocks by channels. A channel is a means of conveying
signals. See Figure 6 .18.
Figure 6.18: Static structure of an SDL system.
A block can be partitioned into (sub) blocks and channels, similarly to the partitioning of the system (a
system can be regarded as a special kind of block, having no channels connected to it from the
environment).
The repeated block-partitioning results to a block tree structure (within the system as the root block).
Partitioned blocks do not contain any process. Leaf-blocks of a block tree structure are not partitioned,
and they contain only processes of different types. Within a leaf-block, signals are conveyed on signal
routes between processes, and between a process and the boundary of the block (a signal route is
analogous to a channel), see Figure 6 .19.
Figure 6.19: Structuring a block into process types.
The statistical structure of an SDL system is reflected in its description. However, a block description
can contain both process descriptions (corresponding to a leaf-block) and a block structure description
(corresponding to a partitioned block). The choice between these two versions of a block must be
made before interpretation time (the result is called consistent portioning subset).
SDL utilises Abstract Data Types (ADT). Most operations on a type can be applied without regarding
the actual implementation. In addition to conventional data types, the language includes specific types
such as time and duration.
Page 45 of 65
IST-2001-34157
6.2 Methodology
The SDL-based methodology will built on the existing formal specifications of standardised IEEE 802
MAC protocol layers. The target will be the formal specification and functional validation of the
PACWOMAN MAC optimisation algorithms, along with an evaluation of the implementation complexity
induced. The adopted SDL-based specification and functional verification methodology consists of the
following discrete phases:
SDL SDL
Requirements SDL Data
Architectural Behaviour
Capture Specification
Specification Specification
Automatic MSC SDL Functional

SDL Analysis SDL Simulations
Genration Verification
MAC Complexity
Evaluation
Figure 6.20: Adopted methodology and mapping on standard SDL development phases.
6.2.1 Requirements Capture
The first task of the SDL design flow is to capture formalises, and analyse requirements for the system
to be developed. The simulation-driven (NS/2, OpNet, etc) high-level definition of MAC optimisations is
going to comprise the starting point for that. MSCs (Message Sequence Charts) graphical notations
are going to be used during this design phase. The service interfaces of the MAC layer of
PACWOMAN MAC layer will need to be sketched first, for defining the service primitives and informal
parameters. A concrete result from the use of early MSCs will be a set of common use cases of the
protocol. These can be later re-used for design verification and simulations.
6.2.2 SDL Specifications Phase
During the architectural and behavioural specification phases, the functional model of the system is
built using the top-down design approach. These phases also correspond to block level and process
level hierarchical layers of the SDL architecture.
The MAC layer specification is started by defining the boundaries (interfaces) between the protocol
and its environment. Next, the specification proceeds by further partitioning the system into SDL
blocks. Dependencies between blocks are defined by connecting them with channels.
The block architecture partitioning will start from the system level and blocks will be subsequently
decomposed into sub-blocks. Channels inside sub-blocks will be also defined during this phase.
Page 46 of 65
IST-2001-34157
Next, the functional behaviour of the MAC protocol is specified by means of active objects (processes
and procedures) and passive objects (data types). Processes describe how signals and the data
exchanged between blocks are handled. Most complex data types, timers, and different kinds of
exceptional behaviour situations will then be defined informally. Procedures with informal
specifications can be effectively used to hide unnecessary details.
6.2.2.1 High-level MAC-layer Architecture

The system level of the generic PACWOMAN MAC-layer model is presented in Figure 6 .20. This is
based on the IEEE 802.11 formal specification and is most appropriate for master-master
communications.
User Interface
MAC_Data_Service_B MAC_Data_Service_B
MDPU_Generation_B
MLME_STA_B
Protocol_Control_B
Transmission_B Reception_B
Channel Access Interface
Figure 6.21: Generic PACWOMAN MAC-layer SDL system-level design
The MAC SDL system consists of the following Blocks:
 MAC_Data_Service_B: This entity validates requests from the user layer and processes &
adds/removes the MAC protocol headers.
 MAC_Management_Service_B: This block implements the MAC MIB (Management Information

Base), handles MIB access by other blocks, and controls filtering and confirmation of MLME
requests.
 MPDU_Generation_B: Implements encryption, fragmentation, and power save queuing.
 MLME_B: This block implements the MAC-sub-layer management entity (MLME). It manages and
monitors the protocol layer, starts the IBSS (Independent Basic Service Set), and controls the
power save state.
Page 47 of 65
IST-2001-34157
 Protocol_Control_B: This entity implements the CSMA/CA based DCF (Distributed Control
Function).
 Transmission_B: This block handles procedures such as FCS (frame check sequence)
generation, timestamp insertion, etc.
 Reception_B: This block accesses channel state (physical and virtual carrier sense), handles IFS
(Inter-Frame Space) slot timing, etc.
6.2.3 SDL Functional Verification
One of the most significant advantages of SDL is that system functionality can be simulated and
verified in early phases of the design. Consequently, errors and misunderstandings are detected and
corrected before proceeding into the implementation phase, which increase the quality of design and
can significantly save time and resources. As in any software design testing is needed for ensuring the
error free operation of an SDL model. The operation of all blocks, processes and data type definitions
must be verified also in exceptional or error situations.
A set of appropriately selected scenarios will be used for verifying the MAC-layer SDL model
functionality. SDL will also be used for the development of test scenarios. This re-use of technology
(SDL used for both system formal specification and validation) will considerably ease and speed up
technical progress.
6.2.4 Complexity Evaluation and Optimisation of the SDL Model
The high abstraction level of SDL leads to a conceptual gap between the formal description and its
physical implementation. Each SDL construct could be translated into a target platform implementation
using a number of ways, each way having a different implementation strategy. SDL contains
properties that separate it from real-world systems:
 SDL description assumes that infinite system resources are available. First, each process
contains an unlimited size FIFO buffer for incoming signals. Second, SDL uses abstract and
infinite size data types. Third, no processing time is assumed for state transitions.
 The operation of SDL processes is parallel, while most systems utilise a single or few processors
for multitasking. In addition, the target languages of the automatic code generation are usually
also sequential.
 Target platforms are usually synchronised while the SDL behaviour is mainly based on
asynchronously communicating processes.
However the properties of an SDL model can lead to conclusions regarding the complexity of the
targeted systems. More specifically, the following SDL attributes are typically used for this purpose:
 The total number of Block instances in an SDL system.

 The number of Process instances in a block type of an SDL System.
 The total number of process instances in an SDL System.
 The total number of States in an SDL system.
 The total number of Timers in an SDL system.
 The total number of Signals in an SDL implementation.
 The total number of Block Types in an SDL system.
Page 48 of 65
IST-2001-34157
 The total number of Signal Lists in an SDL system.

 The total number of Gates in an SDL system.
 The total number of Signal Routes in an SDL system.
 The total number of Channels in an SDL system.
 The total number of Built in Procedures in an SDL system.
 Average Number of Processes per Block Type.
 The average number of states per process type in an SDL system.
 The total number of input events in an SDL system.
 The average number of input events per state in an SDL system.
 The total number of Procedure Calls in an SDL system.
 The total number of Parameters in an SDL implementation.
 The Average Number of Parameters per Built in Procedure in an SDL implementation.
 The total number of output events in an SDL system.
 The average number of gates that are contained in a process type of an SDL system.
 The number of timers that are set in input events of an SDL system.
 The number of signals included in signal lists of an SDL implementation.
 The average number of signals grouped per signal list of the SDL implementation.
 The number of continuous decisions that are nested in the transition area of an input event.
Page 49 of 65
IST-2001-34157
7 CONCLUSIONS
MAC techniques in the context of Wireless Personal Area Networks have been investigated. Starting
with an investigation of the basic principles and challenges from the viewpoint of the PACWOMAN
architecture. A brief review of the involved standards has been presented as well, namely WLAN IEEE
802.11, Bluetooth and WPAN IEEE 802.15.x.
Medium Access Control in the Wireless PAN environment is divided into two fields, one focuses on the
medium and high data rate domain, where ad-hoc routing techniques are applied, and the other
focuses on the low data rate domain, where typically master-slave communication is used. In the
MDR/HDR domain multi-hop connections are considered.
For both theses domains requirements and design constraints are considered and based upon this
design issues for the MAC layer are presented.
For the ad-hoc domain extensions to carrier-sensing techniques have been considered and a receiver
initiated MAC algorithm that applies energy pulses to jam the signal is proposed. More detailed
numerical analysis of the proposed protocol will be presented in D4.3.1.
For the LDR domain the use of the IEEE 802.15.4 protocol will be investigated in order to derive an
adequate parameter set that optimises the overall power consumption. These and other issues, like
for instance the use of simple ALOHA protocols for the ad-hoc option, are going to be studied and
evaluated in the PACWOMAN deliverable D4.3.1. An initial study on link adaptation for UWB systems
was performed, where the issues and parameters were identified to design a adaptable MAC layer for
UWB. This MAC can adjust the physical parameters to meet a predetermined BER.
Several currently available simulation platforms were reviewed and it was concluded to use either
NS/2 or custom-made tools. NS/2 will be used for large system level evaluations, where the whole
protocol stack is simulated and the effects of multiple layers are studied. Custom-made tools are used
to test small algorithms and ideas, before a system level implementation is done.
A SDL specification of the MAC layer of IEEE 802.11 has been initiated to analyse the implementation
complexity of proposed extensions to CSMA MAC. This approach will be used to further analyse the
extensions to the carrier sensing mechanisms from a implementation point of view. In particular
focussing on power consumption and power saving.
Page 50 of 65
IST-2001-34157
APPENDIX A:
MEDIUM ACCESS TECHNIQUES IN ULTRA-WIDEBAND AD HOC
NETWORKS
Conference paper as presented in ETAI 2003 in Ohrid, Macedonia, September 2003.
MEDIUM ACCESS TECHNIQUES IN ULTRA-WIDEBAND AD

HOC NETWORKS
Hiroyuki Yomo, Petar Popovski, Carl Wijting*, István Z. Kovács*,
Nico Deblauwe, Angel F. Baena, and Ramjee Prasad
Department of Communication Technology, Aalborg University, Denmark

{yomo, petarp, carl, istvan, ndeblauw, quillo, prasad}@kom.auc.dk

This paper describes work undertaken partly in the context of the IST-2001-34157 Power aware
Communications for Wireless OptiMised personal Area Network (PACWOMAN). The IST program is partially
funded by the EC.
Page 51 of 65
IST-2001-34157
Abstract--Short-range wireless systems have recently time. IR-UWB meets the requirements for short-range
gained a lot of attention to provide seamless, multimedia wireless systems and additionally, possesses the following
communications around a user-centric concept in so desirable features: mitigated multi-path fading effects,
called wireless personal area networks (WPAN). Ultra- possibility of high bit rates and a unique location ability.
wideband (UWB) technology presents itself as a good Indeed, the IEEE 802.15 Task Group2, which deals with high
candidate for the physical layer (PHY) of WPAN, both for rate PHY for WPAN, has formed a study group (SG3a) to
high and low data rate applications. Although the PHY consider UWB as a candidate for this PHY [3].
issues for UWB in WPAN environments have been
Recently, the PHY issues for UWB in WPAN environments
extensively studied, the problems related to the medium
have been extensively studied [4] [5] [6]. However, the
access techniques remain largely unexplored, especially
problems related to the medium access techniques remain
for targeted WPAN scenarios. In this paper, we consider
largely unexplored, especially regarding the targeted ad hoc
time-hopping spread spectrum (TH-SS) UWB
networking scenarios. In ad hoc networking, the devices are
communication system and discuss how to design medium
interconnected via spontaneously created, disposable
access techniques in order to achieve efficient resource
connections, without relying on a pre-existing infrastructure.
sharing in WPAN environments. We evaluate throughput
These scenarios pose seriously challenging research tasks,
performance by computer simulation in multi-network
since the same medium should be used by many mutually
scenario, and discuss the impact of TH parameters on the
interfering WPANs under the stringent synchronization
throughput performance. Furthermore, we propose a link
conditions imposed by the IR-UWB. The two basic
adaptation mechanism which can adapt TH parameters
modulation schemes used in IR-UWB are Pulse Position
according to interference conditions among collocated
Modulation (PPM) and Bi-Phase Modulation (BPM), while
WPANs. This mechanism aims to maximize the
the two common channelization techniques are Time-
throughput achieved by WPANs while keeping the
Hopping Spread Spectrum (TH-SS) and Direct Sequence
robustness with respect to the basic timing mechanisms
Spread Spectrum (DS-SS). This work analyzes only IR-UWB
used in the TH-SS communications.
communication systems using TH-SS and PPM. Since WPAN
technology uses license-free wireless links, the radio
Index terms—Wireless Personal Area Network (WPAN),
resources should be shared among the collocated and
ad-hoc networking, Ultra-wideband (UWB), Medium
uncoordinated WPANs. Then, devices belonging to a single
Access Control
IR-UWB WPAN should actually share the channel not only
among themselves but also with the devices belonging to the
neighboring, physically collocated networks. In such a
1. INTRODUCTION situation, there are two different levels of multiplexing we
need to achieve. First, the medium access control (MAC)
protocols should coordinate the transmissions of all devices
Short-range wireless systems have recently gained a lot of within a network by using contention-free techniques (e.g.
attention to provide seamless, multimedia communications polling), random-access (e.g. carrier-sensing), or a
around a user-centric concept in so called Wireless Personal combination of both. At the second level, the multiple
Area Networks (WPANs). Salient requirements of these collocated networks must share the resources in a sense to
systems include low-power operation, use of license-free minimize the mutual interference, which can be achieved
wireless links, low cost, scalability and the use of ad hoc through adapting the parameters of TH used within the
networking techniques. Ultra-wideband (UWB) technology networks.
presents itself as a good candidate for the physical layer
(PHY) of such systems, both for high and low data rate In this paper, we consider TH-SS UWB communication
applications. Possible applications are: medical monitoring, system and we show how to design TH parameters in order to
office automation, sensor networks, information services, and achieve efficient resource sharing at the second level of
banking / financial applications [1]. multiplexing, that is, resource sharing among multiple
networks. We evaluate throughput performance by computer
UWB radio communications are an extreme form of spread simulation in multi-network scenario, and discuss the impact
spectrum communication systems, generally defined as of TH parameters on the throughput performance.
operating with a fractional bandwidth greater than 0.25 (i.e. a Furthermore, we propose a link adaptation mechanism that
-3dB bandwidth which is at least 25% of the center frequency can adapt TH parameters according to interference conditions.
used). Alternatively, Federal Communications Commission This mechanism aims to maximize the throughput achieved
(FCC) in USA defines a UWB device as any radio with a by WPANs while keeping the robustness with respect to the
-10dB fractional bandwidth greater than 0.20 or occupying at basic timing mechanisms used in the TH-SS communications.
least 500 MHz of the spectrum [2]. In this paper we will
consider UWB systems based on Impulse Radio (IR-UWB),
which use very short pulses, and have been known for a long
Page 52 of 65
IST-2001-34157
2. ULTRA-WIDEBAND PHYSICAL LAYER The large transmission bandwidth in UWB, in the order of
(UWB PHY) n  100MHz (with n  5 70 ), has as result a higher
immunity to interference effects and improved multipath
fading robustness. The multipath can be resolved down to
2.1. UWB radios differential delays of nanosecond or less (i.e. 30cm or less
spatial resolution). In order to convey the information
UWB radio communications are an extreme form of spread symbols in UWB communications several approaches for the
spectrum communications. The FCC has defined an UWB modulation techniques exist, mostly based on the classical
device as any device with a -10dB fractional bandwidth, base-band modulation types. Modulating the UWB pulse
FBW , greater than 0.20 or occupying at least 500MHz of characteristics such as amplitude (PAM), time position
the spectrum [2]: (PPM), phase (PM), shape (PSM), or any combination of
these, can be used. Another direct consequence of the large
fH  fL communication bandwidth is the possibility to accommodate
FBW  2  0.20 many users, even in multipath environments.
fH  fL
(1) The two most common channelization and multiple access
(MA) techniques in IR-UWB are:
where the upper, f H , and lower, fL frequencies  Direct Sequence Spread Spectrum (DS-SS), similar to
correspond to the -10dB bandwidth. code division multiple access (CDMA) communication
systems but using an UWB radio pulse as chip pulse shape
The FCC also regulated the spectral shape and maximum and,
power spectral density (  -41dBm/MHz) of the UWB
radiation in order to limit the interference with other  Time Hopping Spread Spectrum (TH-SS), which uses
communication systems. The ETSI regulations in EU are UWB pulses pseudo-randomly shifted in time domain.
expected to follow the FCC but with a more restrictive
spectral shape, motivated by a different management of the For simplicity, we restrict our study to TH-SS (SB) IR-UWB
available spectrum [7]. communication systems using binary PPM (BPPM).
Furthermore, we make abstraction of the pulse shape and of
The UWB signals generation methods can be grouped in two the corresponding optimal detection techniques that can be
major categories: used.
 Single-Band (SB) based: employing one single transmis-
sion frequency band, and 2.2. TH-SS UWB description
 Multi-Band (MB) based, employing two or more fre-
quency bands, each with at least 500 MHz bandwidth. For a multi-user (device) scenario, the format of the
(k )
transmitted TH-SS IR-UWB signal, s tx , corresponding to
In the SB solution, the UWB signal is generated using very
short, low duty-cycle, baseband electrical pulses with the k  th user (device) is given by:
appropriate shape and duration. Due to the carrier-less 
characteristics (no sinusoidal carrier to raise the signal to a (k )
s tx (t ( k ) )  w p (t ( k )  jT f  c (jk )Tc  d (j k ) )
certain frequency band) these UWB systems are also referred j  
to as carrier-free or IR-UWB communication systems [6]. (2)

The MB UWB systems can be implemented carrier less where t (k ) is the k  th transmitter's clock time; w p (t )
(different pulse shapes/lengths are used according to the
frequency band) [8] or carrier based (multi-carrier like) [9]. is the used UWB pulse; T f is the time frame allocated for
each UWB pulse; Tc is the time shift step used in
In the IR-UWB solutions, typically the radiated pulse signals
(k )
are generated without the use of local oscillators or mixers, channelization / MA together with the c j TH code
thus potentially a simpler and cheaper construction of the (sequence) allocated to each communication channel/user
transmitter (TX) and receiver (RX) is possible, as compared (device);  is the additional time shift used for M-PPM of
to the conventional narrow-band systems. The characteristics (k )
of the pulse used (shape, duration), determine the bandwidth the information symbols, d j . To increase the reliability of
and spectral shape of the UWB signals. The most common the communication link, the same symbol can be repeated a
pulse shapes used in IR-UWB are: Gaussian monocycle (and certain number of times N p , increasing the processing gain
its derivatives) and Hermitian pulses. The low transmit power
of the system, i.e. d j  const. for j  [ p  p  N p ] and
(k )
levels together with the ultra-fine time resolution of the
system can increase considerably the synchronization any integer p . A TH sequence c j is typically a pseudo-
(k )
acquisition time and the complexity of the receiver.
Page 53 of 65
IST-2001-34157
random sequence with period N p , where each element of WPAN. This level of interference can be controlled by
employing traditional MAC protocols. For instance, in
the sequence is an integer in the range [0  N h  1] . The Bluetooth based WPAN, polling scheme with Master/Slave
time frame T f is divided into equally spaced TH shifts, operation, which is a contention-free MAC technique, has
Tc , so that T f  N h Tc . The TH codes can be chosen been adopted for this coordination [10]. Also, random-access
techniques such as carrier sense multiple access (CSMA), or a
based on their auto- and cross-correlation properties, number
combination of contention-free and random access techniques
of estimated users, etc. similarly to the CDMA systems. The
can be employed. The second level of interference is Inter-
M-PPM is achieved by using the additional time shift  and WPAN Interference, which is the interference among links
Tc  M . In the remaining of this paper we consider only located in different WPANs. Since it is difficult to coordinate
the binary PPM (BPPM) case, thus
(k )
M  2 and d j  0 or the transmissions and completely avoid interference among
devices in these "uncoordinated" WPANs, the multiple
d (jk )  1 . collocated networks must share the resources in a sense to
minimize the mutual interference, which can be achieved by
In a multi-user scenario (ad-hoc networks, etc.) several adopting the transmission technologies that have the inherent
transmitted signals, as given by (2), propagate through the immunity to the interference like SS technologies.
radio channel, interfere with each other and add up at the
input of the receivers of all the users. Additionally, due to the In this paper, we focus on IR-UWB WPAN employing TH-
multipath fading, delay, attenuation and other noise sources in SS transmission as shown in the previous section. Each
the radio channel, the received signals are further distorted. communication channel is defined with single TH sequence,
and one or several TH sequence(s) can be allocated to each
The choice for the systems parameters T f , Tc and  is
WPAN. In each WPAN, the use of the TH sequence(s) is
dependent on the desired system performances (throughput, coordinated either by Master/Slave operation or it is regulated
reliability, etc.) and the radio propagation channel in distributed manner. Assuming that the Intra-WPAN
characteristics (multipath fading, dispersion, etc.). interference is perfectly controlled by some MAC protocol in
In our work we abstract the radio channel propagation each WPAN, in this paper, we consider how to control second
mechanism in terms of signal fading, spread and additional level of interference, that is, Inter-WPAN interference. The
noise. The absolute signal attenuation and delays, introduced only way to control the Inter-WPAN interference is to change
by the different relative geographical locations of the
TH parameters such as N p , N h , Tc according to
users/devices, and the multiple-access pulse collisions are the
only sources of link level errors that are considered and different situations, e.g., number of interfering WPANs. In
analyzed. order to discuss adaptive control of the Inter-WPAN
interference, we have to investigate how these parameters
affect the interference among different links with different TH
3. Medium Access Issues for IR-UWB WPAN sequences, how these parameters interact with each other, and
how the throughput and bit error rate (BER) performance are
influenced by changing these parameters. Therefore, in the
3.1. Medium Access Techniques following section, we investigate the relationship between
network performance and these parameters by simulation.
The role of medium access techniques in wireless networks is
to coordinate transmission access to common radio resources 3.2. Simulation Model and Results
so that the interference among different transmissions is
avoided or decreased and capacity (number of communication
links with satisfied quality of service (QoS)) is maximized in
the network. In the WPAN scenario, we have multiple
collocated WPANs, each of which is composed of personal
devices of a user such as PDAs, laptop PCs, etc. Since WPAN
technology uses license-free wireless links, the radio
resources should be shared among the collocated and
uncoordinated WPANs. Thus, devices belonging to a single
WPAN should actually share the channel not only among
themselves but also with the devices belonging to the
neighboring, physically collocated networks. In the design of
medium access techniques in such a WPAN context, we need
to take account of two different control levels of interference.
The first level of interference is Intra-WPAN Interference
which is the interference among links located in single
Page 54 of 65
IST-2001-34157
The highest throughput is obtained for the case where only 2

Np=10
channels are assumed to be present and N h  2 . With an
x 10
5 PuRe = 10
14
base=2
Nh=2
base=4
Nh=4 increasing number of other networks (channels) the
base=8
Nh=8
12 performance of the N h  2 case degrades. As can be seen
from Figure 1 after 10 networks became active the
10
performance for the case of N h  4 starts to outperform the
case of N h  2 . The better performance of networks
Throughput
8
Throughput
applying a larger N h can be explained from the fact that

6 each WPAN generates less traffic and thus it reduces the total
amount of generated interference. Since the system is in an
4
interference limited domain the performance is improved, in
other words the additionally generated traffic in the case of
2
N h  2 is lost in collisions. This gives an intuition for
possibilities to optimize the UWB system for the number of
0
collocated networks.
2 4 6 8 10 12 14 16 18 20
Number of channels
Number of channels From Figure 2 it can be seen that the BER of the system
increases as the number of collocated WPANs increases. The
Fig. 1 Throughput for and BER can be used as a measure to indicate the number of
To investigate the interactions between the UWB physical 0
Np=10
PuRe = 10
10
layer and the MAC layer a custom-made simulation tool has base=2
Nh=2
base=4
Nh=4
been developed. The simulator investigates the interactions -1
base=8
Nh=8
between collocated WPANs. One WPAN is taken as the 10
WPAN under investigation and the other WPANs are

considered as interfering networks. -2
10
A WPAN is modeled as a cluster of nodes between which the

communication channel is defined by a unique TH sequence.
BER
-3
BER
10
Since here we are interested in the inter-network interference

the internals of a WPAN are not modeled. For the sake of -4
simplicity, we assume that only one TH sequence is allocated 10
to a WPAN. The WPAN is only represented by the TH

sequence, which is used to transmit a constant flow of -5
10
information. We assume perfect coordination of this TH

sequence within a WPAN, which means that there is always -6
10
only one transmission at a time in each WPAN. 2 4 6 8 10 12 14 16 18 20
Number of channels
Number of channels
In accordance with the regulations of the FCC the frequency
band was chosen as f L  3.1 GHz and f H  10 GHz Fig. 2 BER for and
with a center frequency of 6.85 GHz, resulting in an effective
pulse duration of 162.64 ps. The selected modulation method collocated WPANs. Based upon Figure 2 a heuristic for
is Disjoint BPPM. The propagation aspects of the wireless adjusting the value of N h for a fixed target BER can be
channel were modeled using the free space model. The area developed, which can be used to create a reliable packet flow
used in the simulations is a square of 15m. Assuming that the after coding. For example if the target is set to 10 2 then
delay spread of the channel is less than 20 ns, the size of one
bin, containing two pulse positions, was set to be the system can start with applying N h  2 , with 4 or more
2  20  40 ns. The transmitted power in the WPANs is, collocated WPANs the number of bins is increased to 4, and
P  2.5 dBm. The simulation results are averaged over 5 with 12 collocated WPANs the number is further increased to
runs of simulations of more than 12.000 transmitted bits. 8.
Now the performance of the system with a varying number of Maintaining a target BER is investigated further in Figure 3,
collocated WPANs is considered. All WPANs are using the which depicts the number of WPANs that can co-exist while
same parameters. The number of TH shifts or bins within the maintaining a target BER, here the target was set to be
TH sequence of the WPAN is varied from 2 to 8. Figure 1 10 1 . The number of WPANs (channels) is studied as a
depicts the throughput for the case of a pulse repetition of 10. function of the processing gain, parameterized by the number
Page 55 of 65
IST-2001-34157
 Soft link adaptation - The values of N h and N p are

20
changed, but N h  N p  const., such that the transmission
Nh = 2
Nh = 4
Nh = 8
rate is constant, i.e. Tb  const.
15
The hard link adaptation requires explicit exchange of
No. channels for BER < 0.1
information about the spreading sequence among all members

10
of a WPAN, since they should have the same representation
of the current nominal bit rate within the WPAN. Hence, the
hard link adaptation necessarily introduces a communication
5
overhead. The hard link adaptation determines the value of:
Tb

0
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Tc
Np (4)
Fig. 3 Number of maximum channels for a fixed BER for On the other hand, the soft link adaptation can be applied
varying (BER @ and 2-20). autonomously, by each device in the WPAN, during its
transmission and it does not introduce a communication
overhead. Let the minimal value of the hopping base, used in
of bins in a frame. The higher the processing gain N p , the
the system, be fixed at min N h  h0 , such that c j in (2)
(k )
less sensitive the system is to interference, as before also the
number of bins N h can be increased, reducing the number are chosen pseudorandomly and uniformly from
of transmissions and thus the collision (probability). [0  h0  1] . For given  , the maximal value of the
 
processing gain is N p  h0 . For soft link, the transmitter
4. System Design for Interference Mitigation uses chip puncturing by transmitting only N p pulses per
Np
4.1. Hard/Soft Link Adaptation bit, where N p  N p and the value q  N   1 .

p
The discussion in the previous section sets a basis to consider At the conceptual level, the puncturing can be done
the link adaptation in the WPANs based on IR-UWB. In this pseudorandomly: If a pulse should be transmitted in certain
case, the link adaptation is a mechanism to adapt the TH (k )
parameters in order to maximize the throughput achieved by slot, as determined by c j , the WPAN device transmits the
the WPAN. Our proposal for a link adaptation procedures pulse with probability q and with probability 1  q the
considers that the system has a fixed chip duration Tc . This device stays silent in that chip interval. Note that this decision
choice will keep the pulses intact, while all the changes for does not need to be coordinated with the receiver, who will
the actual transmission at the physical layer will be induced perceive the lack of pulse as an erasure in the channel. The

by change of parameters at the MAC/link layer. To start with, average number of pulses transmitted per bit is q  N p .
we assume that the time frame is fully used for chip
To support soft link adaptation, we assume that the devices
transmission, i.e. T f  N hTc . Then, the bit duration can be within a WPAN share a TH sequence defined with a TH base
represented as a function of N h and N p as follows: of size h0 . By using the puncturing rate q  1 , the actual

Tb ( N h , N p )  N h N pTc TH base N h is increased, N h  qN p
 h0 . Alternatively,
(3) during the hard link adaptation, all WPAN devices can adopt
the same value of q , which is autonomously changed
and the nominal bit transmission rate is Rb 
1
Tb . afterwards by each WPAN member via soft link adaptation.
The actual puncturing may be performed according to a
The WPAN communication is adapted to the channel pattern different from the described pseudorandom one.
conditions by varying N h and N p . This variation should
To see the applicability of the hard and soft link adaptation,
be robust with respect to the basic timing mechanisms used in we must consider the interference model for a WPAN.
the TH-SS communication. We define two basic techniques
that can be used in the link adaptation:
4.2. Interference Mitigation
 Hard link adaptation - The nominal bit transmission rate
i.e. Tb is changed;
Page 56 of 65
IST-2001-34157
Since a WPAN uses open access in the unlicensed spectrum, 5. CONCLUSION

we consider that the errors are induced from the interference
with other devices with unlicensed operation, such that the In this paper, we have discussed medium access techniques in
IR-UWB WPAN using TH-SS and PPM. We have evaluated
communication is interference-limited. For the purpose of
BER and throughput performance in multiple WPAN
link adaptation, WPAN differentiates between 2 interferer
scenario by computer simulation and shown that the choice of
types:
TH parameters largely influences the BER and throughput
1. Interference from a collocated non-WPAN entity; performance. Based on these simulation results, we have
proposed a link adaptation mechanism which can adapt TH
2. Interference from identical, collocated WPAN.
parameters according to interference conditions. This
The essential difference between the two is that the mechanism maximizes the throughput achieved by WPANs
interference from a non-WPAN interferer is regarded as while keeping the robustness with respect to the basic timing
statistically invariant. That is, the WPAN considers that the mechanisms used in the TH-SS communications. Our further
interference that the WPAN itself causes to the non-WPAN investigations include the implementation and performance
entity will not initiate change in the transmission pattern of evaluation of the proposed link adaptation mechanism in
the non-WPAN entity. The reliability of the bit transmission dynamic WPAN scenario. Also, the future work includes the
with respect to a non-WPAN interferer can be increased only design of key functionalities for medium access such as
by increasing the number of transmitted pulses per bit, Np. If control channel, beacon channel, synchronization in IR-UWB
the satisfactory value of Np is determined, then there should context.
be no puncturing and a hard link adaptation set the value of
Tb in the WPAN to be: 6. REFERENCES
Tb  N p h0Tc
[1] PACWOMAN Consortium, “State-of-the-Art of the
(5)
WPAN networking paradigm”, Deliverable 5.1,
The situation is different when the interference is induced by PACWOMAN IST-2001-34157, 2002. Available:
collocated, identical WPANs. The interference that a WPAN http://www.imec.be/pacwoman.
is causing to the collocated WPANs is determined by the [2] Federal Communications Commission “Revision of Part
number of pulses transmitted per second, which, for fixed Tb 15 of the Commission's Rules Regarding Ultra-Wideband
qN p Transmission Systems”, First Report and Order, ET
and given q can be expressed as . Each WPAN is
Tb Docket 98-153, FCC 02-48, April 2002.
considered as a selfish agent3 that applies a strategy to [3] IEEE Working Group for WPAN. Available:
maximize its throughput but it is aware that the other WPANs http://www.ieee802.org/15
are also applying such strategies. If all collocated WPANs [4] M. Z. Win and R. A. Scholtz, “Ultra-Wide Bandwidth
increase the pulse repetition rate, the transmission reliability Time-Hopping Spread-Spectrum Impulse Radio for
is not improved. In order to host more collocated WPANs, as Wireless Multiple Access Communications”, IEEE
the results in previous section show, the TH base of each Trans. Commun., Vol. 48, No. 4, pp. 679-691, April
WPAN should be increased. This can be achieved by the 2000.
described chip puncturing.
[5] J. Foerster, E. Green, S. Somayazulu, and D. Leeper,
We have described the mechanisms to apply link adaptation.
“Ultra-wideband technology for short- and medium-
However, one key issue remains - how will the WPAN know
range wireless communications,” Intel Technical Journal,
whether the interference is caused by WPAN or non-WPAN
May 2001, http://developer.intel.com/technology/itj
device? We outline the basic procedure: A WPAN sets the
value of Tb to be low and sets initial value of q  1 . If the [6] M. Z. Win, R.A. Scholtz, "Impulse Radio: How it
works", IEEE Comm. Letters, Vol .2, No.1, pp. 36-38,
BER is not satisfactory, each device tries autonomously to January 1998.
vary q in order to reach higher throughput. If there is
unsatisfactory BER for some predefined timeout, a device in [7] Domenico Porcino, “UWB Regulations \& Coexistence:
the WPAN (e.g. the master) may initiate hard link adaptation From the FCC First Report and Order to the Path to
Approval in Europe”, Tutorial at International Workshop
to increase Tb . The investigation of the optimal interplay
on Ultra Wideband Systems, Oulu, Finland, June 2003.
between the hard/soft link adaptation to maximize the
[8] N. Askar, "Overview of General Atomics PHY proposal
throughput is out of the scope of this paper.
to the IEEE 802.15.3a", IEEE P802.15, March 2003.
[9] J. Joe, "Cellonics UWB Pulse Generators", International
Workshop on Ultra Wideband Systems, Oulu, Finland,
June 2000
3
This terminology is borrowed from the game theory
Page 57 of 65
IST-2001-34157
[10] Bluetooth SIG, Bluetooth Core Specification, ver. 1.1,

February 2001, http://www.bluetooth.com
Page 58 of 65
IST-2001-34157
Page 59 of 65
IST-2001-34157
Page 60 of 65
IST-2001-34157
Page 61 of 65
IST-2001-34157
Page 62 of 65
IST-2001-34157
Page 63 of 65
IST-2001-34157
Page 64 of 65
IST-2001-34157
Deliverable number: D 4.3.2.
Title of Deliverable: Medium Access Control for a wireless ad-hoc environment
Page 65 of 65

WP4.3-CPK-D4.3.2-Medium Access Control For A Wireless Ad-Hoc Environment-13!10!2003-V1.0

Enviado por

Dados do documento

Descrição original:

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

WP4.3-CPK-D4.3.2-Medium Access Control For A Wireless Ad-Hoc Environment-13!10!2003-V1.0

Enviado por

Direitos autorais:

Formatos disponíveis

INFORMATION SOCIETIES TECHNOLOGY (IST) PROGRAMME

Power Aware Communications for

D 4.3.2 - Medium Access Control for Wireless Personal Area Networks

Contractual Date of Delivery to the CEC: August 31, 2003

 Copyrighted by the Participant(s)

DOCUMENT REVISION HISTORY

Version Date Author Summary of main changes

DOCUMENT REVISION HISTORY........................................................................................3

2.1 Basic Principles.........................................................................................................................................12

2.2 Architecture Design..................................................................................................................................13

2.3 Power-saving mechanisms.......................................................................................................................14

2.4 Relevant standards...................................................................................................................................15

3.1 Requirements and Constraints................................................................................................................24

3.2 Design Considerations..............................................................................................................................28

3.3 MAC alternatives......................................................................................................................................29

3.4 Implementation Proposal.........................................................................................................................30

4.1 Requirements and Constraints................................................................................................................35

4.2 Design Considerations..............................................................................................................................36

4.3 MAC Alternatives.....................................................................................................................................37

5.1 Review of possible simulation platforms................................................................................................38

5.2 Simulation Methodology and Scenarios.................................................................................................40

6.1 Description of Approach..........................................................................................................................43

2 MAC TECHNIQUES IN GENERAL

2.1 Basic Principles

Figure 2.1 Overview of the high level PACWOMAN architecture.

NET (TCP/IP) NET (TCP/IP)

LL (IEEE 802.3) LL (IEEE 802.3)

MAC MAC (TDMA, CSMA…)

UWB BT IEEE 802.11 BT H-OFDM

Basic Terminal (bT) Advanced Terminal (aT)

Transmission Power Control

2.4.1.1 The IEEE 802.11 Wireless LAN Architecture

Figure 2.4 The basic Architecture of an IEEE 802.11 Wireless LAN.

2.4.1.2 IEEE 802.11 Topologies

2.4.1.3 IEEE 802.11 Media Access Control

2.4.2.1 Bluetooth Specifications

2.4.2.2 Low power modes

2.4.3 IEEE 802.15.3

- Data Rates: 11, 22, 33, 44, & 55 Mbps.

- Low power consumption

2.4.4 IEEE 802.15.3a

2.4.5 IEEE 802.15.4

- Over air data rates of 250 kb/s, 40 kb/s, and 20 kb/s.

Figure 2.8 The IEEE 802.15.4 WPAN superframe structure.

3 MEDIUM ACCESS CONTROL FOR THE AD-HOC LINK

3.1 Requirements and Constraints

3.1.1 Extensions to CSMA/CA

= Black Burst = Data = Scheduled

Figure 3.9 Basic operation of the Black Burst Scheme.

tsch = trdy - tbslot - tobs - 

3.1.2 MAC layer for H-OFDM

FR i-1 FR i FR2 i+1

Slot (packet, burst) level

Figure 3.11 Illustration of slot, frame and super-frame levels.

3.3 MAC alternatives

1. Every contention has a unique winner

3.4.1 Basic Protocol Operation

Figure 3.13 The proposed scheme - blackburst++.

w  tdata + SIFS + tACK + tmed + bmax + tobs + tCTS + SIFS

3.4.2 Contention resolution

Figure 3.14 Contention resolving - A sample scenario.

Figure 3.15 The proposed scheme - Contention resolution.

Hiroyuki Yomo, Petar Popovski, Carl Wijting, István Z. Kovács,