Escolar Documentos
Profissional Documentos
Cultura Documentos
Gasoline-Engine Management
Emissions-Control Technology
for Gasoline Engines 1 987 722 102 978-3-86522-050-9
Gasoline Fuel-Injection System K-Jetronic 1 987 722 159 978-3-86522-051-6
Gasoline Fuel-Injection System KE-Jetronic 1 987 722 101 978-3-86522-052-3
Gasoline Fuel-Injection System L-Jetronic 1 987 722 160 978-3-86522-053-0
Gasoline Fuel-Injection System Mono-Jetronic 1 987 722 105 978-3-86522-054-7
Ignition Systems for Gasoline Engines 1 987 722 130 978-3-86522-040-0
Gasoline-Engine Management:
Basics and Components 1 987 722 136 978-3-86522-041-7
Gasoline-Engine Management:
Motronic Systems 1 987 722 139 978-3-86522-045-5
Imprint
Automotive Networking
Contents
16 Automotive networking
16 Cross-system functions
17 Requirements for bus systems
19 Classification of bus systems
19 Applications in the vehicle
21 Coupling of networks
21 Examples of networked vehicles
30 Bus systems
30 CAN bus
44 LIN bus
50 Bluetooth
60 MOST bus
71 TTP/C
84 FlexRay
96 Diagnosis interfaces
As electronic systems and components continue to take hold in the automotive sector, so the
number of electronic systems used in motor vehicles is constantly increasing. Such systems
used to operate autonomously to a large extent, but today are combined to form an interconnected
network. Powerful bus systems enable data to be exchanged between these systems.
Networking has taken hold right down to the small-car segment. The engine-management system
exchanges data with other systems which have become standard features (e.g. antilock braking
system, ABS). Many comfort and convenience functions (e.g. power-window units) are likewise
controlled by bus systems. New functions e.g. the further development of ACC (Adaptive Cruise
Control) for stop & go traffic require data from different vehicle domains. In other words, all the
electronically controlled systems in a vehicle must communicate with each other.
Data transfer in the various vehicle domains is subject to different requirements. Various bus
systems have therefore become established. This publication starts by explaining the basic
principles of networking, then provides an overview of automotive networking, and uses several
examples to demonstrate the need for networking in modern vehicles. A further chapter deals with
the most important buses in detail.
UVA0014-1Y
Robert Bosch GmbH
SVA0005Y
Robert Bosch GmbH
Adr 1 Adr 2 Adr 3 Adr 4 In the predictable method the bus access
right is determined before bus access.
Adr 3 It can thereby be ensured that only one
subscriber is using the bus at a time.
Access collisions because of simultaneous
b
bus usage will be prevented if all subscrib-
Id 1 Id 2 Id 2 ers use this method.
Id 4 Id 3
Id 3 Id 3 Id 5
Id 6 Id 5 Id 6
Fig. 8
Id 7
UVA0012-1Y
a Subscriber-oriented
method
Id 3
b Message-oriented
method
Robert Bosch GmbH
5V
Data Data
Bus
User layer line
Application layer E1 E2
T1 T2
Communication layer
Presentation layer
Control Status T Bus level
E
Session layer
0 conducts 0V
Transport layer
1 blocks 5V
Network layer
E1 E2 Bus level
Data link layer 0 0 0V dominant
0 1 0V dominant
Physical layer 1 0 0V dominant
1 1 5V recessive
Physical layer
UVA0027E
SVA0008Y
Physical connection
Robert Bosch GmbH
bus when one node is transmitting a status If no data is being exchanged, the bus level
of 1 and another is transmitting a status is 5 V (microcontroller operating voltage,
of 0. Fig. 11). When the start bit is transmitted
The binary statuses can be depicted in (dominant level), the other station con-
many different ways. The serial interface nected to the bus (receiver) is notified
of the PC, for example, uses +12 V and that a data transfer is starting (Fig. 12).
-12 V, and CAN-B uses voltages of 0 V and The length of the start bit determines a
5 V. The voltages of the serial interface are bit time that represents the basis for the
unsuitable for a bus, since short-circuits entire data transfer. Every subsequent
can occur if several subscribers wish to data bit has the same length. The recipro-
transmit conflicting binary statuses simul- cal of this time corresponds to the data
taneously. transfer rate, i.e. the number of bits that
If the coding allows one level to over- can be transmitted in one second in a
write another, the overwriting level is re- continuous data stream. All participat-
ferred to as dominant, and the subordinate ing stations must be set to the same data
level as recessive. transfer rate.
It is also possible to depict dominant After the start bit has been received,
and recessive levels using visual media. the transmission of an 8-bit data word
A status of 1 (recessive) then corresponds commences (1 byte) with the lowest signif-
to e.g. dark, and a status of 0 (dominant) icant bit (LSB, Low Significant Bit). The re-
corresponds to light. In an optical fiber, ceiver that has synchronized itself to the
an individual node can override all of the start bit scans the data bus between each
others by feeding light into the conductor. data bit and therefore assembles the trans-
ferred data byte.
Bit stream The eighth data bit is followed by the
The application information cannot usu- parity bit. This bit indicates whether the
ally be transmitted directly. In order to number of transmitted ones is odd or even.
make transmission possible, the informa- It therefore allows the receiver to perform
tion is first incorporated as a payload in a simple check for possible transmission
the frame of a message that contains infor- errors. The sequence is completed with
mation to be transmitted. Since all proto- the stop bit, which is placed onto the bus
cols have been developed in accordance
with different requirements, the frame
format differs from protocol to protocol. 12 UART interface transmission frame
1
Priority
2
low
Level
1
1 2 3
0
Time
b Occurrence of events
high
1
Priority
4
3 7
2 6
5
low
Level
1
1 3 4 2 5 6 7
SVA0010Y
Time
Robert Bosch GmbH
0
Max. output
Time slots delay
Time
SVA0011Y
Robert Bosch GmbH
Table 1
Robert Bosch GmbH
Automotive networking
1 Automotive networking
SVA0037Y
Robert Bosch GmbH
30
20
10
SVC0012-1Y
0
W140 W220 W221
(1991) (1998) (2005)
Robert Bosch GmbH
convenience systems than the antilock Hard real-time requirement: the time
brake system (ABS), for example. specification must be strictly adhered
In order to meet these requirements, to. If the specified response time was ex-
mechanisms that detect transmission er- ceeded, the calculated result would not
rors are incorporated in the network pro- be able to be used. This can lead to seri-
tocols. A simple check can be carried out ous problems in safety-critical systems.
using the parity bit, which is calculated in
the transmitter and is transmitted together For example, if time allowances were ex-
with the useful data. This specifies whether ceeded in the ABS system, the incipient
the number of 1's in the transferred byte is locking of the wheels would not be de-
even or off. This information is checked by tected soon enough and the pressure in
the receiver. Single errors can be detected the master cylinder would not be reduced
using this method. in time. This would result in locked wheels.
Another method is the checksum check. The time allowances must also be strictly
If several data bytes are being transmitted, adhered to for many engine-management
the transmitter calculates a checksum system functions. Delays in transmitting
from the individual data bytes using a pre- injection and ignition signals could lead
defined formula and transmits this value. to engine judder and even misfiring.
The receiver also calculates the checksum These reactions must be avoided, since
of the data bytes that have been received they represent a potential danger. Hard
and compares it with the checksum that has real-time requirements must therefore be
been received. If a data transmission error made of these systems.
is detected, the received data is not used However, this does not necessarily mean
and a repeat transmission is requested. that the transmission of data via a bus sys-
tem also has to be subject to these hard
Real-time capability real-time requirements. Adherence to
A real-time system guarantees that its re- soft real-time requirements is usually suf-
sults are calculated within a fixed time in- ficient. If signals from other control units
terval. The duration of the time interval are needed for functions (e.g. a torque re-
depends on the application. The antilock duction request during a shift operation),
brake system (ABS) must react to the the bus system must transmit the data at
incipient locking of a wheel within a few a faster data transfer speed and with a
milliseconds (wheel speed reduction), smaller time delay so that the overall sys-
whereas response times of 100 ms are tem complies with the specified real-time
adequate for actuating the power-window requirements.
motor. Human beings cannot perceive
delay periods of less than 100 ms. Number of network nodes
Different demands are made of real-time The maximum number of nodes to be inte-
behavior depending on the application: grated varies for different areas of vehicle
Soft real-time requirement: the system operation. The number of nodes for com-
generally adheres to the specified re- fort and convenience systems may be high
sponse time, and if these times are occa- due to servomotor networking (e.g. seat
sionally exceeded, it does not produce adjustment) and intelligent sensors (e.g.
any serious effects (e.g. image jerking rain sensors). Several identical busses can
during picture transmission). be used if necessary.
Robert Bosch GmbH
Chassis Drivetrain
Interior Telematics
SVC0013-1Y
Robert Bosch GmbH
4 Gateway structures
a b
Diagnostics
5 Network topologies
a
Sensor CAN
Diagnostics
CAN Drive CAN
Instrument
cluster CAN
LIN
Comfort CAN
LIN
Infotainment CAN
Comfort CAN
LIN
BluetoothTM
b Distance control
CAN
Instrument
cluster CAN
Diagnostics CAN
MOST
Drive CAN
LIN LIN
Fig. 5 LIN
Typical network
topologies for
new-generation
SVA0028E
channel. Up to 15 stereo channels can be driving speed to be read out in the work-
transferred in parallel with the current ver- shop via the connected diagnostic tester
sion of the MOST bus (MOST25 with a total (example: the correct assignment of the
transfer rate of up to 24.8 Mbit/s). wheel-speed sensors must be checked for
the ABS functional test).
Data transfer: examples
The following examples show which sig- Engine speed
nals are measured and evaluated in which The injection and (with a gasoline engine)
systems. the ignition timing are output with a reso-
lution of less than 1 of the crankshaft an-
Driving speed gle. In order to ensure real-time behavior,
The ESP control unit calculates the driving the crankshaft position must be recorded
speed from the wheel-speed sensors. This in the engine control unit. The engine-
variable is transmitted on the CAN-C bus speed sensor scans the crankshaft trigger
(drive CAN). The engine-management sys- wheel and relays the signal to the control
tem needs this value for the cruise control, unit, which calculates both the crankshaft
among other things, and the transmission position and the engine speed. This vari-
control unit determines gear changes from able is used to calculate the injection time
the driving speed. The adaptive speed and the ignition angle, for example.
control (ACC, Adaptive Cruise Control) The engine speed is a variable that is
needs the current driving speed to calcu- needed in many other systems. The engine
late the necessary distance from the ve- control unit therefore outputs it on the
hicle in front and use it as a setpoint value. data bus. The shifting points are defined
A gateway transmits the speed informa- in the transmission control unit depending
tion via another CAN bus (instrument clus- on the speed. The engine speed is needed
ter CAN) to the instrument cluster, which for the ASR function (acceleration slip
displays the value via a needle instrument. control) in the Electronic Stability Program
The CAN-B bus (comfort CAN) is also (ESP) ASR intervention (torque reduc-
connected to the network via the gateway. tion) must not make the engine stall.
Some luxury class vehicles are equipped As in the previous example, the engine
with dynamic seats. The padding of the speed is transmitted to the diagnosis inter-
seats is inflated depending on the speed face and the instrument cluster (display on
and the acceleration, counteracting the rev counter).
the centrifugal force of the driver. This
increases comfort considerably when Turn signaling
cornering. The driver operates the turn-signal lever
The speed information is sent to the (Fig. 6, Item 1). A signal is relayed to the
Infotainment CAN via the gateway and re- steering column control unit via a discrete
layed to the car sound system. This allows line (2) depending on whether the driver
the volume to be adapted to the driving is indicating a right or left turn. This may
speed. The navigation system needs the be a resistance-coded signal, for example.
speed to calculate the position if the GPS The control unit evaluates the signal and
signal is missing (e.g. in a tunnel). detects that the driver is indicating a left
The diagnosis interface is directly con- turn, for example.
nected to the engine and transmission The comfort CAN relays this informa-
control unit via the serial K-line. All other tion to the vehicle power supply control
control units are connected to the diag- unit (3). The indication direction is defined
nosis interface via a virtual K-line that is on the basis of the received information
simulated on the CAN bus. This allows the (normal flash frequency, increased flash
Robert Bosch GmbH
Fig. 6
1 Turn-signal lever
2 Steering column
control unit
3 Electrical-system
6 Data transfer during turn signaling 7 Data transfer during windshield-wiper operation
control unit
4, 5 Turn-signal lights
4
6 Gateway
4 7 Trailer-recognition
3 5 LIN
M control unit
3 8 Turn-signal light
Comfort CAN
on trailer
9 Instrument cluster
2
Comfort CAN
1
Instrument cluster
Fig. 7
CAN 9 1
2 1 Windshield-wiper
6
lever
2 Steering column
control unit
SVA0029E
SVA0030E
7 3 Electrical-system
control unit
8
4 Wiper motor
Robert Bosch GmbH
8 Multimedia networking
4 2
Comfort CAN
MOST
Fig. 8 3
5 1
1 Head unit
2 Radio tuner
3 Amplifier 6 7
SVM0005E
4 CAN/MOST gateway
5 Screen Check commands
6 Control element Multimedia data stream
7 Speakers
Robert Bosch GmbH
Control of an automatic steel folding roof four CAN buses and a further four LIN sub-
Vehicles with automatic steel folding roofs buses. In this example, a total of 13 control
are currently very much in vogue. These units are involved in the control of the au-
roofs are automatically opened and closed. tomatic steel folding roof. The text below
The technical realization of this seemingly describes which functions are performed
simple mechanical function poses a huge by the individual control units to move the
challenge for the networking of the elec- roof.
tronics and control units.
The networking depicted in Figure 9 Roof control unit
shows a typical present-day mid-size ve- Power activation of the motors for the steel
hicle with the additional function of an au- folding roof is performed by this control
tomatic steel folding roof. At its maximum unit. This control unit also assumes the
equipment specification, this vehicle has role of complete monitoring of the move-
over 35 electronic control units (ECUs), ment process. Proximity-type sensors
which communicated with each other via which monitor, record and evaluate the
Instrument cluster
Central OBD socket
(WFS)
gateway
Central
processing unit Electrical-system Door control unit Dynamic headl.
(Radio, RadioNavi, control unit driver leveling contr.
RadioNaviNet,)
Wiper
(master) Door control unit AWD
passenger control unit
Booster AMP Wiper
(slave) Door control unit Transmission
rear left control unit
Voice control RS/LS
Door control unit Airbag
SMLS control unit
rear right
TV tuner (steering column)
Comfort control unit
MF st.w. Selector lever
Telephone (Kessy)
interface box
Roof control unit IR Electric
(CSC) power steering
Telematics
NAR Sounder
PTC heating Steering-angle
NGS sensor
Tire-pressure Seat, driver
monitoring (memory) Brake
control unit
Trailer Seat, passenger Fig. 9
(ABS, ESP, )
control unit (memory) Systems in blue boxes
SVA0047E
ESP
cluster are involved in the
Multifunction
Auxiliary heating control of the steel
control unit
folding roof
Robert Bosch GmbH
movement of the roof are used for this ated. This also applies to the sliding sun-
purpose. In order to inform the driver of roof, which is activated by the roof control
the current status while the roof is being unit.
opened and closed and to prevent any
damage to the vehicle or the surroundings, Trailer control unit
the roof control unit receives numerous The trailer control unit informs the roof
further parameters from other control control unit whether a trailer is hitched. If
units in the vehicle. this is the case, opening and closing of the
roof is disabled.
Instrument cluster
The instrument cluster receives via the Door control units
gateway the current status from the roof The status of the windows and doors is
control unit and from other participating interrogated by all four door control units
control units, and informs the driver ac- and transmitted to the roof control unit.
cordingly. Example: Obstacle behind the Opening or closing of the roof is only
vehicle! Roof movement stopped. begun if all the doors are closed and the
windows are in the correct positions. This
Gateway setting is automatically corrected if the
All the communication requests which are windows are not in the correct positions.
transmitted from one bus system to an-
other are routed via the central gateway. Comfort control unit
The comfort control unit informs the roof
Vehicle power supply control unit control unit of, among other things, the
The vehicle power supply control unit key position. Movement of the roof is en-
checks whether the vehicle battery has abled only if the correct key is inserted.
sufficient charge to enable the roof to be
moved. If necessary, the driver is informed Brake control unit
and no roof movement is performed. The brake control unit uses sensors to
record the wheel speeds and thereby iden-
Parking-aid assistant tify whether the vehicle is moving. The
Before the roof is opened, the parking-aid roof control unit receives the speed infor-
assistant monitors the area behind the mation and enables movement of the roof
vehicle for obstacles to ascertain whether only if the vehicle is stationary.
there is sufficient space available to un-
fold the roof. If an obstacle is detected, Future areas of FlexRay application
the driver is informed and movement Up to now the high-speed CAN bus (CAN
is stopped. In this event, the driver can C) has been used to network control units
decide for him-/herself whether he/she in the drivetrain and in the chassis area. In
wishes to continue moving the roof. In future X-by-wire systems, the mechanical
certain cases the parking-aid assistant can connections, e.g. between steering wheel
detect something which poses no danger and front axle (steer-by-wire) or brake
to the vehicle. pedal and wheel brakes (brake-by-wire),
will be replaced by electrical communica-
A/C control unit tion systems in conjunction with driving-
All windows are automatically closed dynamics control systems.
when the air-recirculation switch is actu-
Robert Bosch GmbH
Furthermore, vehicle architectures have ity Program), AFS (Active Front Steering),
up to new often used several CAN buses and ARC (Active Roll Control). Higher-
which are linked to each other via gate- level driving-dynamics control controls
ways to distribute the high data volume. the individual systems without limiting
Future architectures will use, for a fast, their functionality and prevents negative
powerful connection between several mas- interactions. Figure 10 shows the intercon-
ter computers (which assume, for example, nection of these systems with bus system
central functions in the safety and driver- and sensors.
assistance areas), a backbone bus with a The yaw sensor transmits the vehicles
high data rate to which in each case sub- acceleration values and yaw rate to ESP.
buses (e.g. drivetrain CAN and sensor CAN If a critical driving situation is detected
in the chassis area) are connected. by ESP using these data (e.g. vehicle over-
The new FlexRay bus system satisfies steer), situation-conditioned brake pres-
the demands which will be placed in future sures and engine-management interven-
on the vehicle architecture, such as, for tions are calculated. Higher-level global
example driving-dynamics control also evaluates
high data rates and guaranteed real- the driver-command steering angle from
time capabilities in the drive and chassis the steering-angle sensor and calculates
areas, a supplementary steering angle, which is
large date volume in the backbone, and converted by AFS. More effective and com-
high failure safety of safety-relevant fort-enhancing control can be achieved by
applications (e.g. X-by-wire) this interaction of the individual systems.
In this way, the vehicle can, for example,
through properties such as be stabilized by a corrective steering
high availability and redundancy by movement already at a very early stage
means of two physically independent such that braking interventions can be
channels, partially or even completely avoided.
high data rates with up to 10 Mbit/s per
channel,
data transfer with guaranteed latency,
and 10 FlexRay
synchronicity of all communication
users by means of a global time base. Global driving-dynamics control
Driving ESP
CAN dynamics FlexRay (active
FlexRay is currently still in the develop-
braking)
ment stage, although some initial systems
are already in volume production. The fol-
lowing example is a current approach to a Sensor AFS
technology (Active
FlexRay topology.
Front
Steering)
Excerpt from chassis domain
Concepts for global driving-dynamics Further ARC
compo- (Active
control are being developed to facilitate nents Roll
SVA0048E
Bus systems
high-speed bus
ECU 5 to ECU 11 ECU 9 ECU 10 ECU 11
On the CAN-B
low-speed bus
Robert Bosch GmbH
a V
0
a V
2.0 Differential
SVC0018E
signal
0
Robert Bosch GmbH
a V
CAN_L
5
3.6
Node 1 Node n
1.4 (max. 30)
CAN_H
0
recessive dominant recessive
Time t CAN_H
b V
CAN_H 120V CAN bus line 120V Fig. 4
3.5
a Voltage level of
2.5 CAN_L
CAN_L the low-speed CAN
1.5
SVC0019E
SVC0020E
(CAN-B)
b Voltage level of the
recessive dominant recessive
high-speed CAN
Time t
(CAN-C)
Robert Bosch GmbH
Physical layer
CAN 2.0 B). With 11 bits in the standard
format, it is possible to distinguish be-
tween 2,048 different CAN messages;
Robert Bosch GmbH
in the extended format, this number rises ority (lowest binary value of the identifier)
to over 536 million. is assigned first access, without any data
The advantage of this addressing loss or delay (non-destructive protocol).
method is that the network nodes do not The arbitration principle permits the
require any information about system dominant bits transmitted by a given sta-
configuration and are thus free to operate tion to overwrite the recessive bits of the
fully independently of each other. This re- other stations (Fig. 8). Each station outputs
sults in a highly flexible complete system, the identifier of its message onto the bus
which makes it easier to manage equip- bit by bit, with the most significant bit first.
ment variants. If one of the ECUs requires During this arbitration phase, each station
new information which is already on the wishing to send data compares the level
bus, all it needs to do is call it up from the present on the bus with the level it actually
bus. It is possible to integrate additional possesses. Each station that attempts to
stations into the system (provided they are send a recessive bit but encounters a dom-
receivers) without having to modify the inant bit loses the arbitration process.
existing stations. The station with the lowest identifier, i.e.
the highest priority, makes its way onto
Controlling bus access the bus without having to repeat the mes-
Arbitration phase sage (non-destructive access control).
If the bus is unoccupied (recessive state) The transmitters of lower-priority mes-
and messages are available for sending, sages automatically become recipients of
each station is free to initiate the sending the message just sent by another station.
of its message. The message begins with a They repeat their attempt to send as soon
dominant bit (start-of-frame bit), followed as the bus is free again.
by the identifier. When several stations Without this access control, bus colli-
start to transmit simultaneously, the sys- sions would result in faults. To guarantee
tem responds by employing wired-and unequivocal bus arbitration, therefore,
arbitration (arbiter = logical AND operator) it is not permissible for more than one
to resolve the resulting conflicts over bus node to send a message with the same
access. The message with the highest pri- identifier.
Station 1 Station 3
UAE0284-3E
Message format
The message transfer on the CAN bus is 9 CAN message format
Error handling
If a CAN controller detects a fault or for-
mat error, it interrupts the current trans-
mission by sending an error frame com-
prising six successive dominant bits. This
breaks the stuffing rule that prohibits this
type of bit sequence. If the sender detects
that its message has been interrupted by
an error frame, it stops transmitting and
makes another attempt at a later time.
This effect prevents other stations from
accepting the erroneous message and
thereby ensures consistency of data
across the entire system.
Robert Bosch GmbH
Microcontroller Microcontroller
Output Input
buffer buffer
CAN
controller
Control
unit
System bus interface
Transmission Test for: Control
control unit acceptance, unit
faults
Transmit Receive
buffer buffer
Basic CAN
controller Transmission Test for:
Transmit Receive control unit acceptance,
buffer buffer faults
buffer to which the local computer (appli- Modules without local computer
cation software microcontroller) has ac- A further category of CAN module is one
cess (Fig. 10). Since the buffer capacity that is supplied without a local computer.
is limited, the computer must read the These SLIOs (Serial Linked Input/Output)
received data before new messages are are able to input and output data via ports.
received. Message filtering also takes They are therefore suitable for making
place in this computer. A part of the sensors and actuators bus-compatible at
computer's capacity is therefore used for low cost, but they do need a master that
CAN management. Since the computers controls them.
do not usually have sufficient processing
capacity, modules with basic CAN are pri- Transceiver
marily suitable for low bit rates, or for the The bit stream generated by the CAN con-
transmission of fewer messages but at troller is made up of binary signals. They
higher bit rates. do not yet correspond to the required volt-
The advantage of these modules, in com- age levels of the CAN bus. The CAN-bus in-
parison to modules with full CAN, is the terface module, or transceiver, generates
smaller chip surface and the lower manu- the differential signals CAN_H and CAN_L
facturing costs. and the reference voltage Uref from the
binary data stream.
Full CAN
Full CAN implementation is the protocol of Sleep mode
preference in cases where a station has to The CAN comfort bus must remain ready
manage several messages at high bit rates for operation even with the ignition
and the local computer has no free capac- switched off so that functions such as the
ity for communication tasks. They contain radio, power windows or parking lamp
several communication objects, each of may continue to operate. The bus sub-
which contains the identifier and the data scribers must therefore be supplied by
of a particular message. During the initial- terminal 30 (permanent positive). After
ization of the CAN module by the local terminal 15 has been switched off (ignition
computer, it is decided which messages off), a CAN node may enter sleep mode
the CAN controller should send and which (standby) to relieve the vehicle electrical
received messages it should process fur- system of as much load as possible. The
ther. Received messages are only accepted transmitter part of the transceiver module
(message filtering) if the identifier matches is switched off in this condition to mini-
one of the communication objects. mize the power consumption in this mode
CAN controllers with full-CAN imple- of operation. However, the receiver part
mentation relieve the burden on the local remains active and checks whether mes-
computer by performing all of the commu- sages are being sent on the bus. In this
nication including message filtering in the way, the CAN controller, which also enters
controller (Fig. 11). standby mode, is able to react to a wake-up
The CAN controller can be coupled to message and fully activate the CAN node.
the microcontroller in the electronic con-
trol unit as a stand-alone module by the
address/data bus. Powerful microcon-
trollers have the CAN controller integrated
on-chip. This type of bus coupling is the
more cost-effective and thus the more
common solution.
Robert Bosch GmbH
a b
Engine control unit ABS control unit Instrument cluster
Microcontroller Microcontroller Microcontroller
n n
Rev
counter
output
CAN
area n n
RX TX RX TX RX TX
Transceiver
SVC0024D
SVC0024E
CAN bus
Robert Bosch GmbH
station responds with an acknowledge CAN is also widely used in industrial auto-
(ACK check). mation. These applications are supported
At the second stage, or acceptance layer, by an alliance of companies in the CAN in
the message undergoes message filtering. Automation users group (CiA).
Each station checks whether the received Bosch has concluded contracts with its
identifier is addressed to the particular licensees that guarantee that any CAN im-
station and whether the message is re- plementations will be able to communicate
quired in the application software. If not, with each other. Users will be able to rely
the message is rejected. Otherwise, it on the interaction of any CAN modules.
makes its way to the receive buffer. A flag
notifies the application software that a Characteristics
new message is ready for processing. Standardized in accordance with
The instrument cluster, for example, ISO 11898
calls up the available message, processes Prioritized communication
the engine-speed information and calcu- Data transfer rates: up to 1 MBits/s
lates triggering signals for the actuator of Data capacity: up to 8 bytes per message
the rev counter. Real-time response: the data protocol is
sufficient for the real-time requirements
Standardization in the motor vehicle
The International Organization for Stan- Non-destructive bus-access method
dardization (ISO) and SAE (Society of Au- Low power consumption
tomotive Engineers) have issued CAN stan- Flexibility of configuration
dards for data exchange in automotive ap- Simple and economical design with
plications: twisted line pairs
For low-speed applications up to Very high reliability of data transfer
125 kBit/s: ISO 11 519-2 and 11 898-3 Fault detection and signaling
For high-speed applications faster than Localizing of failed stations
125 kBit/s: ISO 11 898-2 and SAEJ 22 584 Handling of intermittent and permanent
(passenger cars) or SAEJ 1 939 (commer- faults
cial vehicles) Short-circuit resistance
An ISO Standard for diagnosis via CAN The number of nodes is theoretically un-
has also been published as ISO 15 765 limited. However, a limit arises in prac-
tice from the capacitive load of the bus
Standardization makes it possible for and the increasing latencies of messages
components of different manufacturers when a high number need to be sent.
to function together. No adaptations are
required.
Robert Bosch GmbH
LIN bus The LIN bus is suitable for low data rates
of up to 20 kBit/s and is typically limited to
Overview a maximum of 16 bus subscribers.
The increasing use of mechatronic systems The electrical interface can be created
in the motor vehicle gave rise to the idea easily and cost-effectively in the network
of designing a cost-effective bus system nodes. As far as the nodes are concerned,
as an alternative to the low-speed CAN. a distinction is made between the master,
In 1998, several automotive manufacturers which is generally an electronic control
founded a consortium with the aim of de- unit connected to a superordinate bus sys-
veloping a specification for a serial bus for tem, and the slaves. These are intelligent
the networking of sensors and actuators in actuators, intelligent sensors or, quite sim-
the body electronics area. ply, just switches with additional hardware
It was believed that a bus system with for the LIN-bus interface.
simple bus protocol and a simple sequence The bus subscribers are usually ar-
control would make it possible to use even ranged in a linear bus topology and con-
low-capability microcontrollers without nected to each other by a single-wire line.
additional hardware for the communica- This topology, however, is not explicitly
tion interface. specified.
The workgroup's resulting LIN bus spec- Communication on the LIN bus takes
ification was introduced into mass produc- place in a time-synchronous manner,
tion with the Mercedes-Benz SL as early as whereby the master defines the time grid.
2001. Consequently, there arises a strictly deter-
The name, LIN (Local Interconnect ministic LIN bus response.
Network), is derived from the fact that Figure 1 shows an example of a LIN net-
all electronic control units are located work as a subbus in the roof/wiper area of
within a demarcated installation space the motor vehicle. Here, the bus comprises
(e.g. in the door). The LIN, therefore, is a central electronic control unit, as the
a local subsystem for supporting the vehi- master, and the four slaves: mirror, garage-
cle network by means of superordinate door opener, rain/light sensor and wiper
CAN networks. actuator. The master also functions as a
gateway to the Chassis CAN, the Body CAN
and the Diagnostics CAN.
dominant
SVL0002E
0
Time t
Robert Bosch GmbH
LIN frame
Header Response
Synch Synch Ident Data Data Data Data Data Data Data Data Check-
Break Field Field field 0 field 1 field 2 field 3 field 4 field 5 field 6 field 7 sum
SVL0003E
Robert Bosch GmbH
The SynchBreak comprises at least 13 con- Of the 64 possible messages, 32 may only
secutive dominant levels and one recessive contain two data bytes, 16 may only con-
level. tain four data bytes, and the remaining 16
At the end of the synchronization break, eight data bytes each.
the master sends the synchronization field The last bits in the identifier field con-
(SynchField) consisting of the bit sequence tain two checksums, which are used to
01010101. The slaves are then able to ad- check the identifier for transmission er-
just themselves to the time basis of the rors and any resulting incorrect message
master and thus synchronize. assignments.
The synchronization method described
permits a loose specification of the timing Data field
of the bus subscribers. The clocking of Once the header sent by the master node
the master should not deviate more than has been transmitted, it is time for the
0.5 % from the nominal value. The clock- transfer of the actual data to begin. The
ing of the slaves is permitted to deviate by slaves know from the transmitted identi-
up to 15 % before synchronization as long fier whether or not they are being ad-
as the synchronization reduces the devia- dressed and, if they are, they reply with
tion to a maximum of 2 % before the end their response in the data field.
of the message. Several signals can be packed into one
In this way, the slaves can be built with- frame. In this case, each signal has pre-
out an expensive quartz oscillator, e.g. cisely one generator, i.e. it is always writ-
using a simple RC circuit. ten by the same node of the network.
During the data transfer of the bytes,
Identifier it is always the least significant bit (LSB)
The third byte in the header is used as the that is output first. Each byte (8 bits) is
LIN identifier. As with the CAN bus, a con- preceded by a start bit and followed by
tent-based addressing method is used a stop bit, which means that each byte
the identifier therefore provides informa- involves the transmission of ten bits.
tion about the content of a message (e.g. The purpose of the start and stop bits is
engine speed). Based on this information, to resynchronize the nodes and thereby
all nodes connected to the bus decide prevent transmission errors.
whether they would like to receive and The data response of the slaves is veri-
process the message further or simply fied by means of a checksum.
ignore it. This process is known as accep-
tance filtering. LIN description file
Six of the eight bits of the identifier field The configuration of the LIN bus, in other
determine the identifier itself. Their per- words the specification of network sub-
mutations give rise to a possible 64 differ- scribers, signals and frames is managed
ent identifiers (ID). They have the follow- in the LIN description file, or ldf. For this
ing meanings: purpose, the LIN specification provides for
ID = 0 to 59: transmission of signals. an appropriate configuration language.
ID = 60: master request for the com- From the ldf, a set of C codes or header
mands and diagnosis files is automatically generated using suit-
ID = 61: slave response to ID 60 able tools. These are used as the basis for
ID = 62: reserved for manufacturer- implementing the master and slave func-
specific communication tions in the electronic control units con-
ID = 63: reserved for future extensions nected to the bus.
to the protocol
Robert Bosch GmbH
LIN master 1
ECU for air-con-
ditioning system
LIN slave 1
CAN
Sliding sunroof
motor
LIN master 2
SVL0004E
Roof module
1 2 3
Fig. 1
1 Cellular phone
2 Headset 5 4
3 Personal digital
assistant (PDA) FM2 EUROPE 1
SVB0001Y
4 Laptop computer
5 Car sound system
with navigation unit
Robert Bosch GmbH
to more than one piconet, can only func- It is conventional for the transmissions
tion as master in one of these networks. of the master to begin in the even time
A device that is assigned to two picon- slots and those of the slaves to begin
ets in a scatternet must synchronize with in the odd time slots. This characteris-
the network with which it wishes to com- tic conforms to the regular time-divi-
municate. It cannot remain connected to sion multiplexing method (TDD here)
both networks at the same time. Before whereby the master uses one half of the
the device leaves the old piconet, it must time slots and the slaves the other half.
inform its master that it will be unavail- In addition to packets that may occupy
able for a certain period of time. The one time slot, Bluetooth provides for
remaining network subscribers can packets that may occupy three or five
continue to communicate. time slots (multi-slot packets). As soon
A master, too, may leave its own as a master or slave sends a packet three
piconet and become a slave in another. or five time slots long, this transmitter
All communication in its old network, will remain on the same frequency.
however, is broken until it returns and There is no frequency change within
assumes its master function again. a packet (Fig. 6).
A master cannot become master of a Once the packet has been transferred,
second piconet. If this were to happen, the frequency hops as determined by the
both piconets would behave in an identi- hopping sequence (independently of the
cal manner and form a single network. transmission process). In the illustration,
for example, the frequency hops from
Physical data channel fk to fk+3 after the transmission of a 3-slot
A data channel is represented by a hop- packet). The reason for this behavior is
ping sequence between the 79 possible that not every station may have received
frequencies in the ISM band. Any device the transmission and cannot therefore
that is actively participating in a piconet respond specifically to the transfer of
must also hop at the same carrier fre- data in several time slots. For this rea-
quency (frequency fi). As soon as a mas- son, all stations that are not involved in
ter has sent data at frequency fk, a slave the transmission always continue to hop
can respond at frequency fk+1. This pro- in the sequence specified by the master.
cedure is illustrated in Figure 5.
Piconet 1 Piconet 2
M M
S S P Slave
P S 625 ms
Fig. 4
M Master
SVB0004E
SVB0005E
S Slave
P Park mode
SB Standby
Robert Bosch GmbH
1 SEQN
SVB0007E
SVB0006E
8 checksum
Robert Bosch GmbH
Baseband
The LAP and UAP fields form the signifi-
SVB0008E
two devices and also implements security In addition to L2CAP, there are also au-
and authentication functions. dio and control protocols that govern the
The host controller interface (HCI) is handling of audio data and control data.
the interface between the hardware of a Audio applications, for example, can
Bluetooth module and the host-side pro- make direct use of the baseband layer
tocols. The HCI is a hardware abstraction once the audio signals have been coded
and realizes various interfaces for con- accordingly.
trolling the Bluetooth hardware and
transferring data (e.g. via USB or UART). The service discovery protocol (SDP) is
another important host-side protocol.
Host stack It serves to identify and scan for services
The host stack uses the HCI to control with specific properties and to describe
the Bluetooth hardware and transfer services within range of a Bluetooth de-
data. The format of this data is deter- vice. Bluetooth devices are meant to be
mined by the upper protocol layers. The able to interact on an ad hoc basis with
logical link control and adaptation proto- other devices in different environments.
col (L2CAP) implements the abstraction It is therefore necessary to know which
of the hardware's properties, adapts the services are made available by which de-
upper layers of the protocol stack to the vices within range. All devices wishing to
capabilities of the baseband and hides provide services must use an SDP server;
transmission details such as the connec- for all other devices, an SDP client is suf-
tionless or connection-oriented trans- ficient. The SDP prepares the services
mission type. The L2CAP layer essen- available on a device in a service data-
tially has three main functions: base. The service information in the pos-
It can receive packets with a length of session of the SDP server is stored in a
up to 64 kB from the upper layers and service record. The service record con-
decomposes them into smaller data sists of a list with service attributes that
packets (segments) for processing in describe the properties of the service
the lower layers if necessary. At the more precisely and is identified by a
other end, the segments are recom- 32-bit service record handle.
posed back into packets.
It manages the multiplexing and de- Two further protocols in the protocol
multiplexing of several packet sources. stack form the fundamental basis for the
If a packet is being recomposed, the interoperability of Bluetooth devices.
L2CAP layer determines to which pro- The RFCOMM (Radio Frequency Com-
tocol of the upper layers the packet is munication) cable emulation protocol
forwarded. above the L2CAP layer simulates up to
L2CAP offers functions for negotiating 60 virtual serial interfaces derived in ac-
quality of service and configuration pa- cordance with the ETSI 07.10 standard.
rameters. This means, for example, that As a result, almost any software that pre-
the maximum size of the payload can be viously expected a serial interface can
negotiated so that a device with limited work with Bluetooth.
resources would not be overwhelmed TCS BIN (Telephony Control Protocol
by overly large packets. Using the con- Specification Binary) is employed as
figuration parameter for the quality of the means of controlling telephone and
service, it is possible to define the telephony functions. This is a bit-ori-
properties of the data transfer: best ef- ented protocol for establishing voice
fort (best attempt but with no guarantee and data connections between Bluetooth
of data transfer) or guaranteed. devices. The use of appropriate, indus-
Robert Bosch GmbH
try-wide standards has made it possible, The Bluetooth profiles form the hierar-
here too, to ensure wide-ranging com- chy (Fig. 10). The GAP forms the basis
patibility with legacy applications. and describes all the essential functions
that a Bluetooth device must fulfill at the
Applications lowermost level. These include, for ex-
Many more protocols have been adapted ample, the functions for establishing and
for the Bluetooth standard (adapted pro- managing the link, the supported operat-
tocols) and can be found in the protocol ing modes and the security of a link.
stack. Internet applications can, for ex- The SDAP defines the access interface
ample, continue to use TCP/IP through for the service discovery protocol (SDP)
the point-to-point protocol (PPP) or the with which devices can discover or poll
Bluetooth network encapsulation proto- the services offered by other devices.
col (BNEP). For the exchange of vCalen- SDP builds on the GAP. All Bluetooth de-
dars and vCards, it is possible to use the vices must implement these two profiles.
object exchange protocol (OBEX) cov- The SPP is used by most other
ered by the IrDA standard. Bluetooth profiles. One exception to
this is the telephony control profile.
Profiles This is always used when Bluetooth is
The uppermost layer contains the appli- being used for cable emulation or if a
cations and profiles. Profiles represent serial data connection is to be used.
standard solutions for a particular us- The SPP builds on the GAP and uses
age scenario. The Bluetooth specification the RFCOMM protocol.
currently brings together 13 different The GOEP defines the fundamental
applications described as profiles. Each functions necessary for the exchange
profile uses a certain choice of protocol; of complex objects. It defines a client-
in principle, application profiles make server relationship for the exchange of
a different protocol stack available for data. Like the SPP, the GOEP provides
each application. Profiles describe the the foundation for further profiles.
vertical slice through the Bluetooth pro-
tocol stack; the protocols represent the
horizontal layers (Fig. 9). Within the pro-
files, the required and optional functions
of the layers are defined. These stan-
dardized profiles make it possible to en- 9 Profiles and protocols
Profiles
Robert Bosch GmbH
10 Profile hierarchy
SVB0010E
LAN access (LAP) Synchronisation (SP)
On the asynchronous channel, data is Each device has a bypass. When this is
sent in packets. It is therefore suitable closed, the device forwards the signal
for transferring information that does directly and is therefore invisible to the
not require a fixed data rate but does MOST system. This service is useful,
need a high data rate at certain times. for example, when the system is starting
Examples could be the transfer of track and a device requires more time for initial-
information of an mp3 player or a soft- ization, or in the event of a temperature-
ware update. With MOST 25, the asyn- dependent shutoff.
chronous channel has a gross bandwidth There are also other possible structures
of up to 12.7 MBit/s. for a MOST system, e.g. a star topology.
Internally, however, these must always
The available bandwidth can be flexibly be arranged to form a (logical) ring.
distributed between the asynchronous With a star structure, this can be achieved
channel and the synchronous channels through use of a central hub, for example.
by means of the boundary descriptor,
which can also be shifted to the transit Device model
period if relevant preconditions are met. The MOST standard defines the model
shown in Figure 2 for a MOST device for
Topology which the following elements are required:
A MOST system is arranged in a ring struc- At the lowermost level, the physical
ture whereby a device is connected to its layer provides access to the transmission
predecessor or successor in the ring by agent. Departing from the original defini-
an input or output respectively (Fig. 1). tion that provided for optical transmission
One of the devices acts as the timing by POF cable (Plastic Optical Fiber), there
master and generates the data frames are today various physical layers available
for data transfer with which the other with optical and electrical transmission as
devices synchronize. well as various speeds of transfer.
Network service
e
MOST network
Tx
Rx tion 1 ve
Tim sition 2
slav
po ng s
Tim
interface controller
si
i
Rx
Tx
ing
po
MOST physical
la
UVM0001Y
UVM0002Y
layer
Robert Bosch GmbH
2.3 mm
To date, modern vehicles are equipped An area for the control channel (with
exclusively with MOST 25, with POF data MOST 25, two bytes per data frame)
transmission via POF connections. In addi- An area for the synchronous channels,
tion to this, however, further transmission whereby one or more bytes are always
techniques have been defined for the assigned to a specific synchronous
MOST: channel
The optical transmission by glass fiber An area for the asynchronous channel
cable (PCS) with laser diodes (VCSEL),
which offers a greater damping reserve, In the case of MOST 25, the synchronous
supports higher speeds and is less sensi- and asynchronous area demand a 60-byte
tive to high temperatures. share of the data frame. The distribution
An electrical transmission by copper between the synchronous channels and
cable, which is also less sensitive to the asynchronous channel is determined
temperature and is more economical by by the value of the boundary descriptor
comparison, requires additional shield- with a resolution of 4 bytes. The synchro-
ing measures at higher bandwidths. nous area must have at least 24 bytes
This has consequences for costs and (6 stereo channels). This means that
cabling. 24 to 60 bytes are permitted for the
synchronous area and 0 to 36 bytes are
Data transfer permitted for the asynchronous area.
Data transfer on the MOST bus is orga-
nized into data frames, which are gener-
ated by the timing master with a fixed data
rate and passed on by subsequent devices
in the ring.
Data frames
The timing master normally generates
data frames with a cycle of 44.1 kHz,
and in rarer cases 48 kHz. The cycle is
defined by the system manufacturer
(i.e. the vehicle manufacturer in conjunc-
tion with its device suppliers) to suit the
predominant media formats in the system. 4 Data frame
descriptor
1 data frame for MOST 25: 64 bytes
Robert Bosch GmbH
sensors and activates actuators. The Com- Communication Network Interface (CNI)
munication Network Interface (CNI) forms The CNI functions as a temporary firewall
the interface between the host processor because it makes available a memory area
and the controller and provides memory via which data received and to be sent are
areas via which the host processor and the exchanged between the host processor
controller can exchange information. and the controller. This interface repre-
The controller is in the end connected to sents the host processors sole possibility
the TTP/C bus, by way of which it connects of transferring information via the net-
the node with other nodes. The group of work. It is therefore not possible for the
all the nodes connected to a TTP/C bus and host processor to influence the communi-
the bus forms a cluster. cation sequence. In particular, the point in
time at which information is provided by
Host Processor the host processor has no influence on the
The host processor executes the applica- networks communication sequence.
tion, the actual function of the network
node. In order to support the application, TTP/C Controller
an operating system, which also for the The controller (Fig. 2) operates entirely
most part provides the FT-COM layer, runs independently of the host processor. Its
on the host processor as well as the ap- most important components are the proto-
plication. The FT-COM layer introduces a col processor, a local bus guardian, and the
further abstraction level of the technical Message Descriptor List (MEDL).
realization of communication. The applica- The function of the protocol proces-
tion can provide data for transfer via the sor is to prepare information provided
interfaces of the FT-COM layer, thanks to by the host processor in the CNI in such a
which the application designer does not way that they can be transferred as TTP/C
have to deal with deeper concepts, such as frames.
the memory areas in the CNI. The controller is connected via two in-
terfaces with the remaining components
of the network node. The connections to
the host processor and to the transceivers
(drivers) are established via the CNI and
the Logical Line Interface respectively.
1 TTA network
Channel 0
SVT0001E
TTP/C bus
Channel 1
Robert Bosch GmbH
Hostprozessor
3 Message components
Communication Network
Interface (CNI) Application data:
Supplied/processed by the host processor
Protokoll- Messages:
MEDL Data areas interpreted
prozessor
by the host processor
SVT0003E
Transmission Media
Within a TTP/C network, communication
takes place via two channels. The channels
are designated Channel 0 (Ch0) and Chan-
b
Node Node nel 1 (Ch1), or Channel A (ChA) and Chan-
SC nel B (ChB).
Because the TTP/C Specification is
SC formulated very openly with regard to
Node Node the transmission medium, it is possible
to build TTP/C networks using very dif-
ferent media. Both electrical and optical
media can be used here. In particular, it is
c also possible to design one channel to be
Node Node optical and the other to be electrical. Dif-
ferences in the latencies of the different
SC SC
media can be compensated for by way of
Node Node
SC SC the configuration in the MEDL.
The achievable bit rate also varies to-
Node Node gether with this flexibility; however, the bit
rate itself is fixed during operation within
a network and cannot be altered.
It is possible with current implemen-
d tations to achieve bit rates of 5 Mbit/s
Node Node Node for asynchronous data transfer during
system starting and of up to 25 Mbit/s for
SC synchronous data transfer during normal
operation. Prototypes in the laboratory Fig. 4
SC a Bus
environment has achieved bit rates of up
b Star
to 1 Gbit/s.
Node Node Node c Multi-star
d Star/bus
SVT0004Y
combination
SC Coupler
Robert Bosch GmbH
the MEDL.
Cluster cycle
Robert Bosch GmbH
A TDMA slot is fixed at points in time COM layer, also an application, which is
which can be observed using the transmis- processed on the host processor.
sion. But because this is not applicable to The lower layers are implemented in
all the activities of a controller during a the controller, where data communication
slot, a distinction is made between node takes place between the host processor
slot and TDMA slot. and the controller via the CNI interface.
For a node, the slot begins with the PSP. In the controller, the top layer is the Pro-
But because no signal is transmitted in this tocol Service Layer; under this is the Data
phase, external observers lack a precise Link Layer followed at the lowest level by
point in time at which the node slot begins. the Physical Layer.
The first precise point in time which can
be observed from an external source is the Physical Layer
start of the TP. This point is also called the The Physical Layer comprises the require-
Action Time (AT). A TDMA slot therefore ments which are placed on the physical
comprises the time interval which passes transmission of information. Often the
between two ATs. Therefore, there exists connections to be used, the cable types or
in a TDMA slot only the TP and the Inter- the voltage levels are established for this
frame Gap (IFG), in which the PRP, Idle purpose in a protocol specification.
Phase and PSP are combined. Efforts are indeed underway to formu-
A node slot, on the other hand, begins late a specification for a TTP/C physical
with the PSP, which is followed by the TP, layer, but the TTP/C protocol specification
the PRP and the Idle Phase. merely establishes requirements which
must be fulfilled by the physical layer
TTP/C Protocol used. This has the advantage that the pro-
TTP/C systems use the TTA framework tocol is very flexible with regard to the
(Time-Triggered Architecture) as the basis networking and bit rate. It is thus possible,
for which with the exception of the appli- for example, for networking to use both
cation finished solutions can be used. optical and electrical media, which can
The TTA protocol stack (Fig. 7) can be also be combined with each other.
divided hierarchically into three areas. However, the system designer is faced
The top layer is the host layer, the ex- with the problem of initially having to
ecuted application. Under this is the FT- identify and evaluate possible solutions.
Host
Executed application
Layer
Node Dslot Node
slot i-1 Node slot i slot i+1 FT-COM CNI
Redundancy management FT-COM
DTP DPRP DIdle DPSP Message permanence Layer
TTP CNI
PRP Idle PSP TP PRP Idle PSP TP Communication services Protocol
Safety services Service
Action Time
Action Time
(AT)
Jennifer Lundelius-
Cluster Startup of a cluster can run with different levels of
Welch,
All the nodes of a cluster act without syn- precision. Only nodes which have precise
Nancy A. Lynch
chronization during cluster startup; bus clocks should be used for time synchro-
access is therefore not yet controlled via nization. These nodes receive the flag as 2)
TDMA. Because there is always a trans- Master Clocks in the MEDL, whereby Hermann Kopetz,
mission delay caused by the medium the SYF (Synchronization Frame Flag) Wilhelm Ochsenreiter:
used when messages are distributed in a is set for their frames. Nodes with more Clock Synchroniza-
tion in Distributed
network, two unsynchronized nodes of a imprecise internal clocks act as Slave
Real-Time Systems,
cluster can start simultaneously with the Clocks and use the clock information of
IEEE Transactions on
message transmission if the transmission the master clocks. There must be at least Computers, 36(8):
delay is greater than the duration of the four master clocks in a cluster to ensure 933-940, August 1987.
startframe. This can occur during system that clock synchronization is Byzantine-
startup if several nodes are authorized to fault-tolerant. Byzantine faults are faults in
initiate synchronization (coldstart nodes). which different nodes interpret the same
The situation can arise where different transmission differently.
cliques arise within one cluster of which The nodes of a cluster are synchronized
only the nodes of one clique are synchro- by measuring the Action Time (AT) in
nized, but not all the nodes of the cluster. each transmitted slot. If the SYF flag of a
In order to avoid this, all the nodes re- frame is set and the received frame was
ject the first correctly received coldstart faultlessly received, the AT of the frame is
frame. The coldstart nodes wait for a time stored in a stack four deep. In other words,
interval (startup timeout) before trying the last four correction values determined
again to transmit a coldstart frame. Be- are stored in the stack.
cause the size of the startup timeout must The actual clock synchronization takes
be different for all the coldstart nodes in place in frames, for which the ClkSyn
the cluster and differ at least by the size flag (Clock Synchronization) is set in the
of the transmission delay, collisions can MEDL. If this is the case, the mean value
no longer occur during the second cluster is formed from the stack values, where
startup. This process is also known as the highest and lowest values are ignored.
Big Bang. The internal clock is now corrected by the
calculated mean value.
Synchronization TTP/C also supports the inclusion of
In a communication system which controls external reference clocks (e.g. GPS) in the
bus access by means of TDMA, it is impor- system. Time Gateway Hosts are nodes of
tant for the individual users to have as pre- this type which offer an external time ref-
cise an idea as possible of a global time. erence. These nodes transmit the external
Synchronization in a TTP/C system uses correction value as application data in a
for the operation time only the offset cor- frame. The remaining nodes of the cluster
rection by means of which the onset of a store this value in the CNI and include it in
point in time is synchronized to all the in- the calculation of the AT correction values.
ternal clocks of the nodes of a cluster. For
this purpose, TTP/C uses a variant of the
Welch-Lynch algorithm 1) which was de-
scribed by Kopetz and Ochsenreiter 2).
Robert Bosch GmbH
Reintegration of a node
A node must synchronize itself with the N-frame (explicit C-state)
cluster if it has failed due to an incorrect Application data
Header C-state CRC
transmission from the member vector or up to 240 bytes
I-Frame C-state
An I-frame begins with the TTP/C header. The C-state consists of the global time,
This is followed by the current controller information on the MEDL and the member
C-state and finally the CRC value of the vector of the cluster. The MEDL position
frame. corresponding to the Round Slot Position
in the current cluster mode, the current
N-Frame cluster mode and the requested cluster
To ensure that Membership Service can mode change (DMC) is transmitted as in-
function, it is necessary for the C-state to formation on the MEDL.
be included in the frame structure. This An explicit C-state has a size of 96 bits
can occur explicitly or implicitly in the (six words of two bytes each).
case of N-frames.
If the C-State is explicitly transmitted, CRC checksum
the header is followed by the controller TTP/C uses the CRC checksum only to
C-state. This is followed by application identify transmission faults. Transmission
data and then the CRC value of the frame. faults are not corrected. Transmission
If the C-State is implicitly included in the faults are detected by a process whereby
frame, the frame consists of the header, each node determines a CRC checksum for
application data and finally the CRC value the received message and compares this
of the frame, the determination of which with the transmitted checksum.
was initialized with the current C-state. A polynomial is used to determine the
Download frames are a special form checksum. The TTP/C Specification in this
of the N-frame; these are accepted by all respect does not stipulate a specific poly-
the receiving nodes, regardless of cluster nomial; instead, it requires the polynomial
association and C-state. To facilitate this to facilitate a Hamming distance of at
behavior, a constant value which is known least 6. In other words, it must be possible
to all the nodes is used for these frames to within one transmission for at least five in-
determine the CRC. correctly transmitted bits to be detected.
The CRC calculation is initialized by the
Frame Fields Schedule ID, where the ID is split into two
Header parts. One part is used for initialization
The header of all the TTP/C frames is for Channel 0 and the other for Channel 1.
identical. Firstly, one bit signals the frame This prevents a node which does not have
type, whether an I-frame or an N-frame is the correct MEDL or which was connected
transmitted. crosswise to the channels from transmit-
This is followed by three bits, by means ting or receiving successfully.
of which a change of current cluster mode The C-state for initializing the CRC cal-
can be requested. These bits are set to culation is also used in the transmission of
zero if no change has been requested. frames with an implicit C-state. Thus, only
those frames whose CRC checksum was
Application data determined with the same C-state are ac-
The information of the application is trans- cepted as valid.
mitted in this area. The size of this field The calculated CRC value is transmitted
can be determined for each slot of a round at the end of a frame in a 3-byte field.
individually in the MEDL and must not
exceed 240 bytes per slot.
Robert Bosch GmbH
Composability Standardization
Networking technology has changed The development of TTA stretches back
greatly over the years. Where initially over a period of 20 years. Today, develop-
stand-alone systems (e.g. activation of a ment of the Specification is coordinated
turn-signal lamp via switches and relays) by the TTAgroup, a cross-industry consor-
were used, the changeover was quickly tium which was founded by the companies
made to cooperative systems in which Airbus, Audi, Delphi, Honeywell, PSA Peu-
individual components exchange informa- geot Citron, Renault, and TTTech.
tion with each other. The Specification of TTP/C has been in-
However, the interaction of components fluenced in particular by Hermann Kopetz,
is increasingly making it difficult to ensure who between 1979 and 1982 was Profes-
the correct fulfillment of a function. For sor of Computer Process Control at the
this reason, the aspect of composability, Technical University of Berlin and since
i.e. the possibility of combining separately 1982 has run the Institute for Technical
developed subsystems into an overall sys- Information Technology at the Technical
tem without having to verify the function University of Vienna. Professor Kopetz
of the overall system again, is becoming also directed the MARS project and was
increasingly important. substantially involved in each further de-
In order to ensure that this functions, velopment of the protocol.
it is necessary to ensure that a subsystem
autonomously fulfills its functions, but also Characteristics
that this is the case when the subsystem is Support of communication for safety-
integrated in an overall system. relevant functions.
TTP/C facilitates the implementation of Guaranteed transmission properties.
composability through the communication Guaranteed detection of all single faults.
parameters which are stored in the MEDL. Fault-tolerant communication units
(FTU) for fault-tolerant provision of
Verification of the Specification data.
The functionality of TTP/C has been thor- Local and central bus guardians
oughly analyzed. The core algorithms possible.
of the protocol specification have been Low protocol overhead.
formally verified with regard to certain as- Data efficiency of 85 % for transmission.
pects of consistency, stability and safety. Supports composability with regard to
Furthermore, the function of TTP/C sys- time response and ranges of values.
tems has been tested using failure-injec- Functions for the most part formally
tion experiments. This involved the use of verified.
both physical and software faults to check Consistent view by all nodes of the clus-
the fault tolerance and fault-detection ter state thanks to Membership Service
properties. and Clique Avoidance Algorithm.
Clear structuring of the protocol inter-
faces.
No restriction of the bit rate from the
Specification. TTP/C components were
tested with a bit rate of up to 1 Gbit/s.
Experience from product use.
Robert Bosch GmbH
Communication data
Host Configuration data & Communication controller
status information
Synchronization
Communication data
Communication data
signals
Configuration data & Bus guardian
Control signal
Control signal
Control Status
signals signals
Control data &
status information Bus driver
UVF0001E
TxEN_B
RxD_A
RxD_B
RxD_A
RxD_B
TxD_A
TxD_B
CBG CBG
A B Subbus
BD BD
UVF0002E
UVF0003E
measures may then be taken. As a result, the bus driver looks out for the transmis-
the effects of these faults can be confined sion of a wakeup signal, in response to
to small areas of the network. which the node is set to wakeup mode.
4 Wakeup sequence
Node C
non coldstart node power off/reset config ready integration
channel B listen
Channel A wakeup
UVF0004E
pattern
wakeup
B pattern
Robert Bosch GmbH
FlexRay gives the host system the capabil- Coding on the physical layer
ity of influencing all five core mechanisms For coding, FlexRay uses a NRZ method
directly. To arrange this, communication (Non-Return to Zero) in which the two
takes place through an interface the con- identical transmission states are not di-
troller host interface (CHI) between the vided by a range in which the voltage falls
network node and host. back to a zero value. This type of coding
Concentrating on the main interfaces, makes it necessary to have mechanisms
the task areas can be arranged as a proto- with which a network node can divide the
col stack (Fig. 5). At the top level is the ap- transmitted states. To this end, FlexRay
plication, which forwards its commands to adds a sequence behind each transmitted
the CHI. Under the CHI is the POC, which byte: the byte start sequence (BSS). From
in turn has access to the MAC, clock syn- this sequence, each node can detect when
chronization and frame & symbol process- a byte has been transmitted and, with this
ing. In the next lower layer are the pro- information, encode the individual bits
cesses for coding and decoding. The low- of the transmission.
ermost level is represented by the transfer By applying different voltages to the
characteristics of the physical layer. two wires of one channel, it is possible
The bottom three layers are in nodes to create four bus states, which are desig-
that are connected to both channels: nated Idle_LP, Idle, Data_0 and Data_1
double presence because these functions (Fig. 6).
have to be fulfilled for each channel sepa- A bus state is identified by measuring
rately. More precisely speaking, this the differential voltage. Here, the bus volt-
means that for one node connected to age (uBus) is made equal to the difference
both FlexRay channels there is double between uBM and uBP (voltages at BP and
availability of synchronization, MAC, BM).
frame & symbol processing, coding uBus = uBP - uBM
and decoding, and physical layer.
Application V
UVF0006E
0
Physical bus
Time
Robert Bosch GmbH
With this method, the data transfer is pro- Generation of a frame bit stream
tected against external electromagnetic Before a node can transmit a frame con-
interference because these act equally on taining the data of the host, the frame is
both wires and are canceled out in the dif- converted into a bit stream. To this end,
ference. the frame is first decomposed into individ-
Idle_LP (LP = low power) is the state in ual bytes. The start of a frame is populated
which a very low voltage of -200 mV and with a transmission start sequence (TSS)
200 mV is present at BP and BM. This state, followed by a frame start sequence (FSS)
for example, is used to identify the start of (Fig. 7). From the bytes of the frames, an
a transmission. extended byte sequence (EBS) is then gen-
In Idle state, a voltage of 2.5 V with a tol- erated whereby each frame byte is pre-
erance of 500 mV is present at BP and BM. ceded with a byte start sequence (BSS).
To set the channel to Data_0 state, at The 24-bit checksum (CRC) for this
least one transmitting node must apply bit sequence (TSS+FSS+EBS) is now calcu-
a negative differential voltage of -600 mV lated and appended to the bit sequence.
to the channel. To finish the bit stream, a frame end
For the channel to be set to Data_1 state, sequence (FES) is appended.
at least one transmitting node must apply If the frame belongs to the dynamic seg-
a positive differential voltage of 600 mV ment, a dynamic trailing sequence (DTS)
to the channel. can be additionally appended to prevent
If the transmission of an information another node from beginning its transmis-
signal is neither blocked by the bus guard- sion on the channel prematurely.
ian nor the communication controller, a
HIGH bit is signaled by the Data_1 channel Communication cycle
state and a LOW bit by Data_0. Each FlexRay cycle (Fig. 8) contains a
static segment that is transmitted as the
first part of the cycle. The static segment
contains a fixed number of transmission
ranges, the static slots.
Optionally, there may be a dynamic seg-
ment in the FlexRay cycle and this is trans-
ferred in second place. Each dynamic seg-
ment contains a fixed, freely configurable
7 Bit stream (dynamic segment) number of minislots.
Also optional is the symbol window,
MAC MAC which is transmitted as the third element
FSP MAC FSP
of the cycle. It can be used for the trans-
mission of an individual symbol and has
CODEC a b cd the same size as a static slot.
BSS BSS BSS DTS
To terminate the cycle, there is a phase
FSS FES
network idle time (NIT) in which the bus
High
is set to idle state. In general, the length
TxD
Low of the NIT corresponds to the remaining
TSS 1st byte last byte macroticks not used by the static and dy-
sequence sequence namic segment or symbol window. This is
1* gdBit
High not the case if it was detected during syn-
UVF0007E
MSB LSB
TxEN chronization that an offset correction is
Low required through which the length of the
NIT can be increased or reduced.
Robert Bosch GmbH
slots are used that hold the same fixed frame ID m frame ID m+3 frame ID m+5
number of macroticks for all nodes con- t
nected to the network. Macroticks are
frame ID m+3 frame ID m+7
ranges in which exactly one frame can
m m+1 m+2 m+3 m+4 m+5 m+6 m+7 m+8
be transmitted.
Dynamic slots (transmission windows Slot counter Dynamic slot Dynamic slot
channel B with without
of variable size) are now created based on transmission transmission
the minislots. The maximum size of a dy-
Dynamic segment containing
namic slot is limited by a fixed number of
UVF0009E
gNumberOfMinislots minislots
minislots specified during configuration. Transmission may only start within the first
pLatestTx minislots of the dynamic segment
8 Communication cycle
t
Communication
cycle level
static segment dynamic segment symbol window network
idle time
Arbitration
grid level
static slot static slot minislot minislot
action point action point action point
Macrotick
level
macrotick
Microtick
UVF0008E
level
microtick
Robert Bosch GmbH
gdMacrotick
Microtick 0 1 2 3 4 5
UVF0010E
level
vMicrotick pdMicrotick
Robert Bosch GmbH
Welch, Lynch,
A New Fault-Tolerant
Algorithm for Clock
Synchronisation,
Information and
Computation, vol.77,
No.1, April 1988
Robert Bosch GmbH
1 ISO 9141-2
ISO 14230-1
SAE J1850
ISO 11898 b Customer-specific
ISO 14230-4 ISO 15765-4
communication
ISO 15765: 1-3 (CAN) b
This standard reproduces the existing Manufacturer-specific
Layer
Layers of the
protocols, ISO 14230: 1-4, on the CAN. K line: KWP 2000 CAN / UDS
ISO 15765-3 OSI reference model
It defines a way to leave untouched the 7 ISO 14230-3
ISO 14229-1 7 Application
fundamental format of messages ex- 6 6 Presentation
changed between tester and electronic 5 ISO 15765-3
5 Session
1 Physical
2 ISO 14230-2 ISO 11898-1
CAN bus
K line
UBat
Diagnosis interface
Tester
SVA0017E
SVA0019E
Tester
Robert Bosch GmbH
Header Initialization
The header is made up of no more than To be able to send messages, each elec-
four bytes. tronic control unit must support a subset
The format byte (Fmt) contains informa- of the possible message formats. At the
tion about the form and composition of beginning of communication, the format
the message. Two bits (A0 and A1) pro- supported by the electronic control unit
vide address information, while six bits is communicated to the tester by means
(L0 to L5) provide the quantity of data of key bytes as part of initialization.
bytes. With this coding, there can be up The ISO standards specify which of the
to 63 data bytes. If L0 to L5 are set to possible initialization procedures are cus-
zero, the number of payload bytes must tomary for ECU diagnostics. The following
be coded using the length byte (Len). initialization procedures are supported:
The target address (Tgt) indicates the 5-baud initialization and
communication partner for which the Fast initialization
information is destined.
The source address (Src) indicates the
communication partner that sent the
information.
The length byte (Len) specifies the quan-
tity of payload bytes (max. 255 bytes).
Up to a value of 63, the coding can take
place in Fmt. For more than 63, the 5 Message format of the KWP 2000 protocol
Checksum calculation
Data bytes
Up to 255 bytes of payload can be sent per
message. The first byte of the payload is
always a service-identification byte (SId). A1 A0 L5 L4 L3 L2 L1 L0
The subsequent data bytes are data that
differ depending on the service con-
cerned. 0 0 No address information
0 1 Exception mode (CARB)
1 0 With address information,
physical addressing
SVA0020E
These different initialization methods are recognition pattern. From the bit sequence
used either for CARB diagnostics or for 01010101 (binary format), the tester de-
after-sales diagnostics. Furthermore, a termines the baud rate of the electronic
distinction can be made between physical control unit, which may be between 1,200
initialization and functional initialization, and 10,400 baud. The electronic control
which introduces the topic of point-to- unit then transmits the two key bytes,
point or point-to-multi communication. which communicate the header format
With functional initialization, a group of and timing method that the electronic
electronic control units is addressed and control unit supports. As acknowledgment
initialized. A function initialization can that communication has been successfully
only be successful if all the electronic con- established, the tester sends the comple-
trol units of this group support the same ment of the second key byte back to the
baud rate, the same transmission timing electronic control unit and, in return, re-
and the same protocol. By contrast, only ceives from the electronic control unit the
one electronic control unit is initialized in complement of the address. Initialization
the case of physical initialization. is then complete and the regular data
transfer may begin.
5-baud initialization
Communication is initiated by means of Fast initialization
a 5-baud address sent by the tester. It is Fast initialization establishes communica-
thus possible for each electronic control tion in a shorter time than 5-baud initial-
unit to be addressed individually and ad- ization. With fast initialization, the tester
dressing is clearly separate from the data sends a wakeup pattern (WuP) comprising
transfer. a low and high phase, each lasting 25 ms
The tester outputs the address onto the (Fig. 7). At the end of the WuP, the tester
K line at a transfer rate of 5 baud (Fig. 6). sends the StartCommunication Service at a
10 bits (eight data bits, one start and one rate of 10,400 baud. The electronic control
stop bit) take approx. 2 seconds. unit returns a positive reply containing the
On receiving the address, the electronic key bytes. Initialization is then complete
control unit responds with the baud-rate and regular communication may begin.
SVA0022E
The first frame contains the PCI, an ad- (Separation Time) specifies the interval
ditional length byte (DL, Data Length) and to be maintained between consecutive
the first six data bytes (Fig. 11). In the frames.
high nibble, the PCI contains the informa- The flow-control frame is followed by
tion for identifying the frames as the first the consecutive frames. Again, the PCI
frame. The extended data length (XDL) in the high nibble contains the identifier,
is stored in the low nibble. Together with while the low nibble contains the sequence
the DL, a 12-bit data word is formed with number (SN). With the first consecutive
which it is possible to express values from frame to be sent, the SN is set to 1. The SN
0 to 4,095. is incremented with each subsequent con-
The transmitter sends the first frame secutive frame; after 15, the SN restarts
and waits for an acknowledgment of re- from 0. By evaluating the sequence num-
ceipt from the receiver in the form of a ber, the receiver is able to detect whether
flow-control frame. This message again or not all the frames have arrived.
contains the PCI, the high nibble of which
identifies the frame as the flow-control
frame. The low nibble contains the flow
status (FS), which can be used to authorize
or delay the sending of further frames
(consecutive frames).
11 Format of a segmented message
7 6 5 4 3 2 1 0
10 Diagram of message transfer
0 0 0 1 XDL
First frame
a Transmitter Receiver
3 2 1 0 7 6 5 4 3 2 1 0
XDL DL
Single frame
b 0 1 2 3 4 5 6 7
PCI BS STmin Data
b Transmitter Receiver
First frame 7 6 5 4 3 2 1 0
0 0 1 1 FS
Flow control Flow control
Consecutive
fram e Clear to send
Fig. 10
STmin Consecutive 0 0 0 0 (CTS)
frame Block a Message transfer
0 0 0 1 Wait (WT),
size (BS) receiver in a single frame
Consecutive
frame not ready (unsegmented
c 0 1 2 3 4 5 6 7 message)
Flow control PCI Data b Segmented
message transfer
Consecutive
frame
Last block 7 6 5 4 3 2 1 0
SVA0026E
SVA0025E
Consecutive Fig. 11
frame ( BS) 0 0 1 0 SN a First frame
Consecutive frame
b Flow control
c Consecutive frame
Robert Bosch GmbH
a b
Fig. 1
a ABS8
(current system)
b ABS1 from 1970
(transistor-
SVA0038D
technology device,
did not go into
series production)
Robert Bosch GmbH
the engine management as to how a TCS- New systems, such as automatic ranging
requested reduction of the drive torque or parking-aid assistant for example, have
was effected (air path, fuel path or igni- their own ECU with the determining sen-
tion path). It was therefore necessary to sors and actuators. However, they increas-
come up with a more powerful interface ingly rely on being supplied with further
via which a desired torque and a dynamic- information by the network. Thus, the ESP
response request could be transmitted. control unit (Electronic Stability Program)
By contrast, the actual torque, the engine supplies the network with the information
speed and the current setting reserve were on the vehicle speed. The vehicle radio can
to be transmitted to the TCS control unit. It use this information, for example, to adapt
proved complex and expensive in terms of the volume to the vehicle speed.
the number of cables required to transfer To keep on top of the complexity, it is
these different data via discrete and, for necessary when developing new systems
example, pulse-width modulated inter- to specify jointly with the creator all the
faces. The CAN bus system (Controller Area variables obtained from the network. This
Network) was introduced in 1991 as an relates in particular to signal quality and
alternative to discrete cabling. In this way, signal availability. The theme of in-vehicle
the foundations for the modern networking safety always play an important role here.
of systems in motor vehicles were laid. The receiving system provides a limp-
home function if the signals obtained from
Technology of the present day the network are unavailable.
In todays mid-size and luxury-class Because of the powerful networking
vehicles, virtually all the ECUs are net- between the ECUs, a good many new per-
worked directly or indirectly (e.g. via gate- formance features can even be achieved
ways) with each other (Fig. 2). Networking completely without additional hardware,
goes so far to some extent that 60 or more i.e. purely by means of data communica-
ECUs communicate via several CAN buses tion and software. One example of this is
and further communication systems, such the opening of the side windows through
as MOST (Media Oriented Systems Trans- longer actuation of the radio remote con-
port) or LIN (Local Interconnect Network), trol for the central-locking system. Thus,
communicate with each other. for example, the vehicle can be uniformly
Diagnostics CAN
Sensor CAN
Instrument
Central cluster CAN
gateway
Drive CAN
LIN
Comfort CAN
LIN
SVA0039E
MOST
Robert Bosch GmbH
ventilated in the summer months when an important task when defining new
the doors are opened. The power-window networking concepts.
units and the central-locking system ex- Thus, for example, the logic circuit
change the necessary information. The for the finger-protection function of the
software required for this purpose runs power-window units is located in many a
either on the ECU for the central-locking design variations directly in the ECU on
system or the ECU for the power-window the power-window motor. The activation
units. In many vehicles the two systems signal for normal operation, e.g. the men-
share a common ECU so that new soft- tioned window opening by radio remote
ware-based performance features can be control, is transmitted via a LIN bus from a
integrated even more easily. central ECU of the body electronics (body
This demonstrates a trend which is al- computer). A client-server architecture is
ready encountered in body electronics: referred to in this respect. Finger protec-
the integration of individual ECUs to form tion is a local function which for reasons of
central ECUs (Fig. 3). These central ECUs safety and on account of its time require-
are connected with the sensors and actua- ments cannot usually be realized on a
tor either via discrete, analog cables or distributed basis.
via buses. The latter reduce significantly
the number of pins in the ECU plug and Development trends
thereby also reduce the cabling costs. Sen- The above-mentioned centralization and
sors and actuators connected via buses are the use of intelligent sensors and actua-
also known as intelligent sensors and tors in the field of body electronics are
actuators. These must for the purpose of expected to extend in the coming vehicle
bus connection have on board electronic generations to the other electronics fields.
circuitry, which in many cases also con- Domain master computers will be used
tains the sensor-signal conditioning or ac- to perform central functions in the areas of
tuator driver functions. At the same time, comfort and convenience, safety, driver as-
however, the use of electronic circuitry sistance, infotainment, and energy supply
gives rise to higher costs in the sensors or (Fig. 4). The ECUs of the intelligent sensors
actuators. Minimizing the overall costs of and actuators distributed in the vehicle
electronics and cabling thus represents are dependent on these master computers.
a b
Fig. 3
1 Mirror 1 3 6 8
4 9
2 Door ECU
3 Power-window unit
4 Check unit 2 LIN
5 Lock
5 10
6 Mirror (LIN)
7 Central ECU (to the other doors) (to the other doors)
CAN
8 Power-window unit
SVA0040E
CAN
(LIN)
7
9 Check unit (LIN)
10 Lock (LIN)
Robert Bosch GmbH
Body & Comfort Infotainment & Cockpit Vehicle Motion Passive Safety Power Train TV Television
Electronics Electronics Electronics Electronics Electronics
VDU Vehicle Dynamics
Unit
Robert Bosch GmbH
Driver
.
dSteering wheel dSum cVehicle
Vehicle
Variable steering
ratio
dSuperposition
Further interference
compensation
Stabilization Actuator
control
.
Fig. 5
cSetpoint
SVA0042E
Signal-flow diagram
showing active front Vehicle model
steering as an example
Robert Bosch GmbH
automotive industry has up to now failed in the form of ECUs, the first step is to look
to agree to a standardized functional for synergy with the further technology
structure of the vehicle. However, many blocks to be integrated. A technological
promising approaches are contained network is created (Fig. 6). If, for example,
in the AUTOSAR Initiative, to which a specific sensor technology is available
Bosch is actively contributing ideas from for an active-chain link whose signal is
CARTRONIC. required by another active chain, this will
also be used. This occurs even if this sen-
Technology model, sor is overspecified for the additional user.
Technological network It is nevertheless important to store the
The technology model describes what original requirement in a database as this
technical realization is used for the speci- synergy may no longer be present in an-
fied function blocks without already com- other vehicle.
bining these into modules, such as e.g. The automotive industry usually
electronic control units (ECUs). Technol- uses the nomenclature e.g. according to
ogy blocks are created. DIN 19 227, IEC 617-4 and other standards
Thus, the signal filtering function can to describe the hardware.
be realized with discrete components by
means of a digital circuit or filter software Integration models
on a microprocessor. Even a controller Node model
function can be executed with a discrete The links of the technological active chains
electronic circuit or a microprocessor. are combined into groups at different loca-
Voltage stabilization can be achieved by tions, the nodes. Here, strict adherence
either a smoothing capacitor or an elec- to the optimum cost of integrating the
tronic DC/DC converter. technology blocks is maintained. Thus, at-
The decision as to which realization tempts are made, for example, to integrate
technology to opt for is dictated on the the software parts of several technological
one hand by the function and on the other active chains on a common microproces-
hand by the costs. Another important sor. Sensor signals are used repeatedly
dictating factor is reuse. Before the tech- where possible and actuators used com-
nology blocks are combined into modules monly where possible. But the history
Software
7 ECU model
Sensor mC Actuator
LIN-bus driver
Robert Bosch GmbH
ECU
Firmware
AUTOSAR Runtime Environment (RTE)
Standard
Software
API 2 Standardized
VFB & RTE Standardized Standardized AUTOSAR AUTOSAR
AUTOSAR
relevant Interface Interface Interface Interface
Interface
ECU
API 1 Services Communication
Abstraction
RTE Standardized Standardized Standardized
Standardized
Operating Complex
API Application
API 0 System Device
Standardized Drivers Program Interface
Interface ECU Electronic Control
API 3 Private
Basic Software
Microcontroller Unit
Interfaces inside Abstraction
Basic Software RTE Real Time
SVA0045Y
UAS1038-1Y
tailed three-dimensional models are used 1 2 3 4 5 6 7 8
F Fuse
P6 Clock/timer
in the later development phase.
W1 Vehicle antenna
X18 Diagnosis socket
30 30
15 15
S1
P1
012
57a 15 15x 50
S38
G1
B+ D+ 50 30
G M
3
Fig. 9 U
G2 Alternator with
P1 P1 T1
voltage regulator
UAS1030-1Y
M1 Starter motor
P1 Display 31 31
1 2 3 4 5 67 8 9 10 11
S1 Ignition switch
T1 Ignition coil
Robert Bosch GmbH
E/E development process Such a restriction can be, for example, the
The E/E development process links the space available in the center console for
individual draft stages with each other on installing ECUs. Another restriction can be
a logic and time basis and provides quality the maximum permissible heat dissipation
criteria at the beginning and the end of a in a location which influences the power
draft stage. electronics positioned there. Thus, for
Because E/E architecture for automotive example, the audio amplifier in vehicles is
applications is still a young discipline, the frequently installed in the luggage-com-
processes at automobile manufacturers partment areas since the heat in the cock-
and suppliers still differ greatly. This re- pit area cannot be adequately dissipated.
lates both to the number and sequence of
the draft stages and to the quality criteria. Evaluation of models
The following must be observed for all ap-
Requirement management proaches: During the transition from one
The requirements decisively determine model hierarchy to the next (e.g. from the
the decisions of the E/E architect. It is ad- function model to the technology model),
visable to distinguish between functional a list of evaluation criteria (e.g. reuse or
and non-functional requirements. Func- testability) is compared with a portfolio of
tional requirements refer to the desired specimen solutions (e.g. bus technologies).
performance features when the vehicle is Evaluation of the specimen solutions using
being used. Non-functional requirements the criteria allows a solution to take shape
refer to the technical solution and are on the basis of the purely functional re-
therefore also known as draft restrictions. quirements and irrefutable boundary con-
ditions (MUST criteria). This procedure
is also known as QFD (Quality Function
11 Example of a 2D space model Deployment).
An alternative procedure consists in
Lights Lights
front left front right comparing a reference solution (e.g. the
Engine compartment previous networking model) using the
Engine compartment evaluation criteria with alternative solu-
tions. This does indeed deliver fast results,
Engine Engine
comp. comp. but possibly not the global optimum.
inside left inside right
Because the evaluation criteria are gen-
Mirror Interior f. m. Mirror
f. le. Interior Interior f. ri. erally weighted differently by the automo-
f. le. f. ri. bile manufacturers, the electronic systems
Door f. le. Door f. ri. of the vehicles sometimes differ consider-
Roof
Door Door ably from each other.
Interior
Interior Interior
Door r. le. r. ri. Door
Luggage
compartment interior
Luggage
compartment
Luggage
compartment
SVA0046E
Lights Lights
rear left rear right
Robert Bosch GmbH
M S
Macrotick, 92 Scatternet, 53
Master-slave, 9 Single-wire line, 32
McMess, 98 Sleep mode, 88
Membership Service, 80 SOS faults, 74
Mesh topology, 7 Space model, 112
Message Descriptor List, 73 Star topology, 5
Microtick, 92 Startup, 88
Minislots, 91 Static segment, 91
Models, EE architecture, 108 Symbol window, 92
Modulation method, Bluetooth, 52
MOST application layer, 67 T
MOST bus, 60 Technological network, 109
Multimaster, 9 Technology model, 109
Multiplex applications, 20 Time-Triggered Architecture, 71
Timer control, 13
N Time synchronization, 93
NetBlock, MOST, 66 Transmission agent, MOST, 62
Networking, 4 TTP/C, 71
Network management, LIN, 48 TTP/C Controller, 72
Network master, MOST, 66 TTP/C Network, 74
Network model, 111 TTP/C Protocol, 77
Network nodes, 31 Two-wire line, 32
Network organization, 8
Network service, MOST, 62 W
Network topology, 4 Wakeup, 88
Node model, 109
O
Offset correction, 93
OSI reference model, 10
P
Physical layer, 10
Piconet, 53
POF cables, 62
Power classes, Bluetooth, 52
Protocol layers, 34
Protocol services, TTP/C, 79
R
Real-time applications, 19
Real-time capability, 18
Ring topology, 6
Robert Bosch GmbH
Abbreviations A E
ABC: Active Body Control EBS: Extended Byte Sequence
ABS: Antilock Braking System ECU: Electronic Control Unit
ACC: Adaptive Cruise Control EDC: Electronic Diesel Control
ADC: Analog Digital Converter EDR: Enhanced Data Rate
AMA: Active Member Address EEM: Electric Energy Management
AMP: Audio Amplifier EMC: Electromagnetic Compatibility
AMS: Application Message Service EPB: Electric Parking Brake
ARS: Active Roll Stabilizer EPS: Electric Power Steering
ASC: Active Suspension Control ESD: Electrostatic Discharge
TCS: Traction Control System ESP: Electronic Stability Program
AT: Action Time ETC: Electronic Transmission Control
AUTOSAR: Automotive Open Systems
Architecture F
FEC: Forward Error Correction
B FES: Frame End Sequence
BCM: Body Computer FOT: Fiber Optic Transceiver
BD: Bus Driver FSK: Frequency Shift Keying
BG: Bus Guardian FSR: Force Sensitive Resistance
BGE: Bus Guardian Enable FSS: Frame Start Sequence
Bit: Binary Digit FTDMA: Flexible Time Division
BM: Battery Management Multiple Access
BM: Bus Minus FTU: Fault Tolerant Unit
BNEP: Bluetooth Network FTM: Fault Tolerant Mean Value
Encapsulation Protocol
BP: Bus Plus G
BSS: Byte Start Sequence GAP: Generic Access Profile
GFSK: Gaussian Frequency Shift
C Keying
CAC: Channel Access Code GOEP: Generic Object Exchange
CAN: Controller Area Network Profile
CAS: Collision Avoidance Symbol GPS: Global Positioning System
CBG: Central Bus Guardian
CDC: Continuous Damping Control H
CDM: Code Division Multiplex HCI: Host Controller Interface
CDMA: Code Division Multiple HDL: Hardware Description Language
Access HFM: Hot-film air-mass meter
CGW: Central Gateway (German: Heifilm-
CHI: Controller Host Interface Luftmassenmesser)
CMS: Control Message Service HHC: Hill Hold Control
CNI: Communication Network HUD: Headup Display
Interface
CPM: Clear Pending Mode Change I
CRC: Cyclic Redundancy Checksum I/O Ports: In/Out Ports
IAC: Inquiry Access Code
D IC: Integrated Circui
DAB: Digital Audio Broadcasting IEEE: Institute of Electrical
DAC: Device Access Code and Electronics Engineers
DI: Direct Injection IFG: Interframe Gap
DLC: Differential Locks Control INIC: Intelligent Network Interface
DMA: Direct-Memory-I/O-Access Controller
DMC: Deferred Pending Mode Change
Robert Bosch GmbH
Abbreviations | 119