Você está na página 1de 6

IPASJ International Journal of Electronics & Communication (IIJEC)

Web Site: http://www.ipasj.org/IIJEC/IIJEC.htm


A Publisher for Research Motivatin........ Email: editoriijec@ipasj.org
Volume 2, Issue 4, April 2014 ISSN 2321-5984


Volume 2, Issue 4, April 2014 Page 9


Abstract
In this study, the performance of multiple input multiple-output channel estimation methods are studied using training
sequences. The popular linear least squares (LS) and minimum mean-square-error(MMSE) approaches are considered as basic
channel estimators. Further, the new scaled LS (SLS) and relaxed MMSE techniques are also studied as these require less
knowledge of the channel second-order statistics and/or have better performance over the conventional LS and MMSE channel
estimators. The MIMO channel estimation can be further improved by directly optimizing channel states such that the cost
function is minimized using Differential Evolution optimization technique.

Keywords: OFDM-MIMO, Channel estimation methods, Differential Evolution (DE) algorithm

1. Introduction
BECAUSE of the growing demand for high data rates in wireless communication systems, array-based transceivers and
space diversity methods have recently become an intensive area of research. It has been shown both analytically and
using field tests that in rich scattering environments, multiple input multiple-output (MIMO) techniques can greatly
increase the capacity of wireless systems. However, to use the advantages that MIMO systems can offer, accurate
channel state information (CSI) is required at the transmitter and/or receiver. For example, the performance of transmit
beam forming is entirely determined by the accuracy of the CSI at the transmitter. If space-time coding is used, then the
availability of an accurate CSI at the receiver is crucial for the performance of space-time decoders. Therefore, accurate
channel estimation plays a key role in MIMO communications.
One of the most popular and widely used approaches to the MIMO channel estimation is to employ pilot signals (also
referred to as training sequences) and then to estimate the channel based on the received data and the knowledge of
training symbols. Recently, there has been a growing interest in training-based MIMO channel estimation. In, the
problem of training sequence design for MIMO channels has been linked with the channel capacity. There are several
works where different training schemes are considered for both the flat-fading and frequency-selective MIMO cases.
For example, in, the maximum likelihood (ML) method has been considered for BLAST training, and orthogonal pilot
signals have been investigated. Further study of this estimator is reported in. In, a least squares (LS) training-based
channel estimation technique for orthogonal frequency-division multiplexing systems with multiple transmit antennas
is developed. In, the conventional LS channel estimation is improved using the minimum mean-square-error (MMSE)
symbol estimate, and training design issues are discussed. In, optimal choices of training signals are investigated in the
case of multiple transmit antennas and single receive antenna in application to several training-based channel
estimation schemes, including the linear LS and MMSE estimators. In, a pilot symbol aided modulation is used to
estimate doubly selective fading channels, and an MMSE-based training scheme with orthogonal training is
considered. In, a general discussion on optimal MIMO training schemes is given based on the LS criterion. Blind and
semiblind MIMO channel estimation techniques are discussed in as alternatives to the training-based channel
estimation.
In this paper, we extend the results of to the MIMO case and study the performance of training-based flat block-fading
MIMO channel estimation. Four training-based channel estimators are considered, which offer different tradeoffs in
terms of performance and a priori required knowledge of the channel second-order statistics.
First, the traditional LS method is considered, which does not require any knowledge about the channel parameters.
Then, a refined version of the LS estimator is developed, which is referred to as the scaled LS (SLS) estimator. It is
shown that the proposed SLS estimator offers a substantially improved performance relative to the LS method but
requires that the ratio of the trace of a specifically defined matrix of channel correlations and the receiver noise power
be known a priori.
Then, the linear MMSE channel estimator is studied. The latter technique is shown to be able to outperform both the
LS and SLS estimators, but it requires the full a priori knowledge of the aforementioned matrix of channel correlations
and the receiver noise power.
Development of Channel Estimation methods
and optimization

P.Prannoy Chakravarthi
1
, Kala Praveen.B
2


1
M-Tech Communications Engineering Student,Vellore Institute of Technolgy,Vellore,TN,India
2
Associate Professor,Vellore Institute of Technolgy,Vellore,TN,India
IPASJ International Journal of Electronics & Communication (IIJEC)
Web Site: http://www.ipasj.org/IIJEC/IIJEC.htm
A Publisher for Research Motivatin........ Email: editoriijec@ipasj.org
Volume 2, Issue 4, April 2014 ISSN 2321-5984


Volume 2, Issue 4, April 2014 Page 10

Finally, the relaxed MMSE (RMMSE) method is introduced, which represents a simplified and approximate version of
the linear MMSE method that requires only the knowledge of the trace of the matrix of channel correlations and the
receiver noise power.
DE algorithm is a stochastic optimization method minimizing an objective function that can model the problem's
objectives while incorporating constraints. The algorithm mainly has three advantages; finding the true global
minimum regardless of the initial parameter values, fast convergence, and using a few control parameters. Being
simple, fast, easy to use, very easily adaptable for integer and discrete optimization, quite effective in nonlinear
constraint optimization including penalty functions and useful for optimizing multi-modal search spaces are the other
important features of DE.
Moreover, optimization of the function is achieved by using the DE algorithm which improved the channel estimation.

2. BACKGROUND
Let us consider flat block-fading MIMO systems with t transmit and r receive antennas. The r x 1 complex received
signal vector can be expressed as
s
i
= H p
i
+ v
i
(1)
where H is the r x t complex random channel matrix, p
i
is the t x 1 complex vector of the transmitted signals, and v
i
is
the r x 1 complex zero-mean white noise vector.
It should be stressed that in any statistical expectation below, the matrix H is treated as random. At the same time, any
estimator H of is supposed to obtain an estimate of a particular realization of this random matrix that corresponds to
the current block of the received data.
In order to estimate the channel matrix H, let N t training signal vectors

be transmitted. The
corresponding r X N matrix S=[ ] of the received signals can be expressed as

S = H P +V (2)
Where
P =[ ]
is the t X N training matrix and V=[ ] is the r X N matrix of sensor noise.

The task of a channel estimation algorithm is to recover the channel matrix H based on the knowledge of S and P .

3. Channel Estimation Methods
3.1 LS CHANNEL ESTIMATOR
Knowing P and received data, the realization of the channel matrix can be estimated using the LS approach as

(3)
where is the pseudoinverse of P and ( . ) denotes the Hermitian transpose. We will use the
following transmitted training power constraint:
(4)
Where is a given constant value and is the Frobenius matrix norm.

3.2. SCALED LS CHANNEL ESTIMATION
The LS estimate does not necessarily lead to the estimate of with the lowest MSE. Therefore, it is meaningful to
optimally scale the LS channel estimate to further reduce the channel estimation error. Scaled LS estimators that
further reduce the MSE by allowing for a bias is a common approach that has emerged in statistics.
Assuming that the values of tr{R
H
} and are given in advance and we obtain that the SLS channel estimate can be
written as


(5)

IPASJ International Journal of Electronics & Communication (IIJEC)
Web Site: http://www.ipasj.org/IIJEC/IIJEC.htm
A Publisher for Research Motivatin........ Email: editoriijec@ipasj.org
Volume 2, Issue 4, April 2014 ISSN 2321-5984


Volume 2, Issue 4, April 2014 Page 11

Note that the SLS estimator is a function of the ratio . Therefore, this ratio has to be known (or preliminary
estimated) when using the SLS approach.

3.3.MMSE CHANNEL ESTIMATION
Let us obtain a linear estimator that minimizes the estimation MSE of H. It can be expressed in the following general
form:

MMSE
= SA
0
(6)

where A
0
has to be obtained so that the MSE is minimized.

(7)

Hence, the linear MMSE estimator of H can be written as

(8)


3.4 RMMSE CHANNEL ESTIMATION
The MMSE channel estimator assumes the perfect knowledge of the matrix . This assumption may be unrealistic in
practical applications. Therefore, we relax it and simplify the MMSE estimator by using the matrix in lieu of in
where the parameter is adjusted to minimize the MSE.

We have that for any training matrix, the RMMSE channel estimator is given by
(9)

4. Differential Evolution Algorithm
The DE algorithm is a population based algorithm like genetic algorithms using the similar operators; crossover,
mutation and selection. The main difference in constructing better solutions is that genetic algorithms rely on crossover
while DE relies on mutation operation. This main operation is based on the differences of randomly sampled pairs of
solutions in the population.

The algorithm uses mutation operation as a search mechanism and selection operation to direct the search toward the
prospective regions in the search space. The DE algorithm also uses a non-uniform crossover that can take child vector
parameters from one parent more often than it does from others. By using the components of the existing population
members to construct trial vectors, the recombination (crossover) operator efficiently shuffles information about
successful combinations, enabling the search for a better solution space.

An optimization task consisting of D parameters can be represented by a D-dimensional vector. In DE, a population of
NP solution vectors is randomly created at the start. This population is successfully improved by applying mutation,
crossover and selection operators.

The main steps of the DE algorithm is given below:

Initialization
Evaluation
Repeat
Mutation
Recombination
Evaluation
Selection
Until (termination criteria are met)
IPASJ International Journal of Electronics & Communication (IIJEC)
Web Site: http://www.ipasj.org/IIJEC/IIJEC.htm
A Publisher for Research Motivatin........ Email: editoriijec@ipasj.org
Volume 2, Issue 4, April 2014 ISSN 2321-5984


Volume 2, Issue 4, April 2014 Page 12



The evolution function is optimized by using this algorithm.
The Evolution Function =
5. Stimulation Results
For the MIMO-OFDM model a 4X4 antenna model is considered. By considering the no. of OFDM symbols as 128 and
no. of frames as 100.

Table.1 Parameters used for stimulation












Fig .1 The Comparison plot of all Channel estimation methods


Fig .2 The frequency plot of all Channel estimation methods
Parameter Values
Modulation Technique BPSK
No.of frames 100
No.of Receivers 4
No.of Subcarriers 128
No.of Training vectors 5/10/15
No.of Transmitters 4
Optimization Technique Differential Evolution
IPASJ International Journal of Electronics & Communication (IIJEC)
Web Site: http://www.ipasj.org/IIJEC/IIJEC.htm
A Publisher for Research Motivatin........ Email: editoriijec@ipasj.org
Volume 2, Issue 4, April 2014 ISSN 2321-5984


Volume 2, Issue 4, April 2014 Page 13


Fig .3 Comparison of Optimized estimation and MMSE,RMMSE


Fig .4 Comparison of DE algorithm by varying no.of generations


Fig .5 Comparison of DE algorithm by varying population size

6. Conclusion
The accuracy of the channel estimation techniques is improved by minimizing the cost function using the differential
evolution approach. The DE algorithm based optimization technique has less computational complexity which reduced
the time elapsed for the optimization of the evolution function. Further the Channel can be estimated accurately which
allows us to eliminate the noise added due to transmission through the channel which improves the voice clarity
through any channel.

IPASJ International Journal of Electronics & Communication (IIJEC)
Web Site: http://www.ipasj.org/IIJEC/IIJEC.htm
A Publisher for Research Motivatin........ Email: editoriijec@ipasj.org
Volume 2, Issue 4, April 2014 ISSN 2321-5984


Volume 2, Issue 4, April 2014 Page 14

References
[1] K.Vidhya and Dr.K.R.Shankar Kumar, CHANNEL ESTIMATION OF MIMO-OFDM USING CUCKOO
SEARCH ALGORITHM in J ournal of Theoretical and Applied Information Technology 20th September 2013.
Vol. 55 No.2
[2] M.P. Chitra and S.K. Srivatsa, "BER Analysis of Coded and Uncoded MIMO-OFDM System in Wireless
Communication", Indian Journal of Computer Science and Engineering, Vol. 1, No. 4, Dec 2010.
[3] Neetu Sood, Ajay K Sharma and Moin Uddin, "On Channel Estimation of OFDM-BPSK and -QPSK over
Nakagami-m Fading Channels", An International Journal (SPIJ), Vol. 4, pp: 239-246, 2010.
[4] M. Biguesh and A. B. Gershman, Training-based MIMO channel estimation: a study of estimator tradeoffs and
optimal training signals, in IEEE Transactions on Signal Processing, Vol. 54, NO. 3, MARCH 2006.
[5] Marcello Cicerone, Osvaldo Simeone and Umberto Spagnolini, "Channel Estimation for MIMO-OFDM Systems
by Modal Analysis/Filtering", IEEE Transactions on Communications, Vol. 54, No. 11, Nov 2006

Você também pode gostar