International Journal of Scientific & Engineering Research, Volume 4, Issue 5, May-2013 1851

ISSN 2229-5518

A Survey on Channel Estimation Techniques in

MIMO-OFDM Mobile Communication Systems
R.S.Ganesh, Dr. J.Jaya Kumari

Abstract— A Modern wireless broadband system of MIMO-OFDM (multiple input multiple output- orthogonal frequency division
multiplexing) is more popularbecause of good data transmission rate and its robustness against multipath fading & good spectral efficiency.
This system provides reliable communication & wide coverage. A main challenge to MIMO-OFDM system is retrieval of the channel
state information (CSI)accurately and synchronization between the transmitter & receiver. The channel state information is retrived with
the help of various f estimation algorithms such as training based, blind and semi blind channel Estimation. This paper describes the basic
introduction of OFDM, MIMO-OFDM system and explains the different channel estimation algorithms,optimization techniques and their
utilization in MIMO system for 4G wireless mobile communication systems.

Index Terms— Channel Estimation, Channel State information,LS Estimation,MMSE Estimation, MIMO-OFDM,Pilot Carriers, Mean
Square Error, Spectral Efficiency.

——————————  ——————————

1 INTRODUCTION

F OURTH Generation Mobile system (4G) has very good

features than previous generation networks such as 2G &
3G .data transmission speed is verh high when compared
2 ORTHOGONAL FREQUENCY DIVISON

IJSER
with previous generation mobile systems. It can fully supports MULTIPLEXING
multimedia services with extreme quality, audio, video files, 2.1 Overview of OFDM
wireless internet and other broadband services with superior Orthogonal frequency-division multiplexing (OFDM) is a type
quality. This technology provides the user to select any desired of frequency-division multiplexing (FDM) method which can
service with more freedom & flexibility be used as a digital multi-carrier modulation technique. Usual-
Mobile communication systems transmit information by ly a large number of closely-spaced orthogonal sub-carriers
changing the amplitude or phase of radio waves. In the are used to carry data. The data is splited into various parallel
receiving side of mobile system, amplitude or phase can vary data streams or channels, one for each sub-carrier. Each sub-
widely. This causes degradation in the quality of system since carrier is modulated by digital modulation techniques such as
the performance of receiver is highly dependent on the quadrature amplitude modulation (QAM) or Quadrature
accuracy of estimated instantaneous channel. In a wireless phase-shift keying (QPSK) at a low symbol rate, maintaining
link, channel state information (CSI)provides the known total data rates similar to conventional single-carrier modula-
channel properties of the link. It provides the detail of signal tion schemes in the same bandwidth. The Modullator outputs
propagation between transmitter and the receiver and tells are lcombined and the resulting signal is transmitted. It
about the effects of scattering, fading. The CSI can could be upconverted and amplified if needed. This scheme is
incorporate current channel conditions with transmission mostly used in various applications such as digital TV &
data for achieving reliable communication . This CSI should be audio broadcasting, wireless LANs, Wi-Fi, WiMAX, LTE,
estimated at the receiver and fed back to the transmitter. ultra-wideband (UMB) systems..
The channel state information can be obtained through dif-
ferent types of channel estimation algorithms .This estimation 2.2 Features
can be done with a set of well known sequence of unique bits In OFDM, the sub-carriers are orthogonal to each other. It
for a particular transmitter and the same can be repeated in avoids the interference between the sub-channels and hence
every transmission burst. Thus the channel estimator no need of guard bands. Therefore the design of both the
estimates the channel impulse response for each burst transmitter and receiver becomes easy .Unlike conventional
separately from the well known transmitted bits and corre- FDM, a separate filter is not necessary for each sub-channel.
sponding received samples.This paper describes the funda- The orthogonality also allows high spectral efficiency. But
mentals of MIMO-OFDM system and study of various channel OFDM requires accurate frequency synchronization between
estimation techniques and their performance. the receiver and the transmitter
It is more easy to transmit a large number of low-rate da-
————————————————
ta streams in parallel instead of a single high-rate stream. It is
easy to insert a guard interval between the OFDM symbols if
• R.S.Ganesh is currently working as Associate Professor in Elcctronics and
Communication Engineering in Noorul IslamUniversity, India, E-mail: the symbol duration is high.By this way , the inter symbol in-
r_s_ganesh@rediffmail.com terference is eliminated. The guard interval also eliminates the
• Dr.J.Jaya kumari is currently working as Professor &Head in Elcctronics and necessity of pulse-shaping filter.
Communication Engineering in Noorul IslamUniversity, India
,E-mail: hellojayakumari@rediffmail.com
IJSER © 2013
http://www.ijser.org
 International Journal of Scientific & Engineering Research, Volume 4, Issue 5, May-2013 1852
ISSN 2229-5518

2.3 Mathematical Description Precoding is one of the multi-stream beam forming tech-
If N sub-carriers are used, and each sub-carrier is modulated nique which is employed at at the transmitter. In beam form-
by M alternative symbols, the OFDM symbol alphabet con- ing, the same type of signal is emitted from each one of the
sists of MN combined symbols. transmit antennas with appropriate phase weighting such that
the maximum received input signal power at the receiver .
The lowpass equivalent of OFDM signal is given as This technique increases the received signal gain, by employ-
ing signals emitted from multiple antennas and also reduces
𝑁−1 the multipath fading effects. It requires exact knowledge of
𝑗2𝜋𝑘𝑡�
𝑣(𝑡) = � 𝑋𝑘 𝑒 𝑇 ,0 ≤ 𝑡 < 𝑇 (1) channel state information (CSI) at the transmitter.
𝑘=0
Spatial multiplexing requires MIMO antenna configuration.
In spatial multiplexing, a high rate signal is splited into sevar-
Where 𝑋𝑘 - Data Symbols al low rate data streams and each stream is transmitted with
the help of different transmit antennas which are having the
N- Number of Subcarriers
same frequency . If these signals arrive at the receiver antenna
T- OFDM Symbol Time. array with different spatial signatures, the receiver can easily
separate this stream of data into parallel channels. It is one of
3 MIMO-OFDM SYSTEM the excellent technique to increase the channel capacity and
improves high signal-to-noise ratios (SNR). The maximum
MIMO-OFDM (multiple input multiple output- orthogonal
number of spatial streams is limited by less number of
frequency division multiplexing) is a modern wireless broad
transmitting antennas at the transmitter or at the receiver.
band technology which has great capability of high rate data
Spatial multiplexing can be used with or without the
transmission and its robustness against multi-path fading and
knowledge of transmitter.
other channel impairments.
Diversity Coding techniques are used when there is no
In MIMO system, multiple number of transmitters at one

IJSER
channel knowledge at the transmitter. In diversity coding a
end and multiple number of receivers at the other end are ef-
single data stream is transmitted with a coding technique
fectively combined to improve the channel capacity of wire-
called as space-time coding. The signal is emitted from each
less system. This technology highly improves the spectrum
of the transmit antennas with full or near orthogonal coding.
efficiency, reliability of system & coverage area . A simple
Diversity coding exploits the independent fading in the multi-
MIMO system with two transmit antennas and two receiving
ple antenna links to enhance signal diversity. Because there is
antennas shown in figure 1.
no channel knowledge, there is no beam forming or array gain
from diversity coding. Spatial multiplexing can also be com-
bined with precoding when the channel is known at the
h11 transmitter or combined with diversity coding when decoding
s y reliability is in trade-off.
1 h12 1 Spatial multiplexing techniques makes the receivers very
complex.Therefore it is usually combined with Orthogonal
frequency-division multiplexing (OFDM) or with Orthogonal
s . y Frequency Division Multiple Access (OFDMA) modulation,
Userdata . Userdata
where the problems created by multi-path channel are han-
2 2 stream
stream dled efficiently.
. MIMO technology is one of the major attracting technique
. . . in wireless communications, because it offers significant in-
. Channel . creases in data throughput and coverage without additional
Matrix H bandwidth or transmit power. It also has high spectral effi-
ciency and link reliability or diversity . Because of these prop-
sM yM erties, MIMO is an important part of modern wireless com-
munication standards such as IEEE 802.11n (Wi-Fi), 4G, 3GPP
Long Term Evolution, Wi-MAX and HSPA+.

3.1 Mathematical Description

s y A narrowband flat-fading channel with multiple transmit
Transmitted vector Received vec- and receive antennas (MIMO) ,the system is modeled as
tor
y=Hx + n (2)
Fig. 1 MIMO System Model
where and are the receive and transmit vectors, respec-
MIMO can be sub-divided into three main parts such as pre- tively, and H,n are the channel matrix and the noise vector
coding, spatial multiplexing and diversity coding. respectively.
IJSER © 2013
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 4, Issue 5, May-2013 1853
ISSN 2229-5518

Ideally, the channel matrix is known perfectly. Due to

channel estimation errors, the channel information can be rep-
resented as
𝑣𝑒𝑐(𝐻)~𝐶𝑁(𝑣𝑒𝑐(𝐻𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒 ), 𝑅𝑒𝑟𝑟𝑜𝑟 ) (3)

Where 𝐻𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒 is the channel estimate and 𝑅𝑒𝑟𝑟𝑜𝑟 is the es-

R

timation error covariance matrix.CN is the circular symmetric

complex normal.

3.2 Channel Estimation

In a wireless communication link , channel state information
(CSI)provides the known channel properties of the link. This
CSI should be estimated at the receiver and usually fed back to
the transmitter. Therefore, the transmitter and receiver can
have different CSI. The Channel State information may be in-
stantaneous or statistical. In Instantaneous CSI, the current
channel conditions are known, which can be viewed by know-
ing the impulse response of the transmitted sequence. But Sta-
tistical CSI contains the statistical characteristics such as fad-
ing distribution, channel gain, spatial correlation etc. The CSI
acquisition is practically limited by how fast the channel con- Fig.2 Classification of channel estimation algorithms
ditions are changing.
In fast fading systems where channel conditions vary rap- The Semi-blind channel estimation algorithm is a hybrid

IJSER
idly under the transmission of a single information symbol, combination of blind channel estimation and training based
only statistical CSI is reasonable. But, in slow fading systems channel estimation which utilizes pilot carriers and other nat-
instantaneous CSI can be estimated with reasonable accuracy. ural constraints to perform channel estimation.
So channel estimation technique is introduced to improve ac-
curacy of the received signal.
The radio channels in mobile communication systems are 4 CHANNEL ESTIMATION ALGORITHMS
usually multi path fading channels, which are causing inter Siavash M. Alamouti [1] proposed a simple two-branch
symbol interference (ISI) in the received signal. To remove ISI transmit diversity scheme. In this scheme two transmit anten-
from the signal, many kind of detection algorithms are used at nas and one receive antenna provides the same diversity order
the receiver side. These detectors should have the knowledge as maximal-ratio receiver combining (MRRC) with one
on channel impulse response(CIR) which can be provided by transmit antenna, and two receive antennas. This diversity
separate channel estimator. scheme does not require any bandwidth expansion and its
3.3 Classification of Channel Estimation computation complexity is similar to MRRC. The BER per-
formance of this transmit diversity scheme was compared
Basic clasifcation of channel estimation algorithm is shown in
with MRRC scheme. Also the implementation issues such as
figure .2.They are training based, blind channel estimation or
power requirements, delay effect, channel estimation errors
semiblind chanel estimation.
were discussed.
The training-based channel estimation can be performed by
Eric Pierre Simon et. all[2] presented a novel pilot aided al-
either block type pilots or comb type pilots. In block type pilot
gorithm for MIMO-OFDM system operating in a fast time
estimation, pilot tones are inserted into all frequency bins
varying environment. This joint estimation algorithm jointly
within the periodic intervals of OFDM blocks. This estimation
estimates multipath Rayleigh channel complex ampli-
is suitable for slow fading channels. But in comb type pilot
tude(CA) and the carrier frequency offset(CFO).Time varying
estimation, pilot tones are inserted into each OFDM symbol
CA is approximated by Base Expansion Model (BEM)and CFO
with a specific period of frequency bins. This type of channel
parameters are modeled by first order AR process. This algo-
estimation is very much suitable where the changes even in
rithm is better in fading channel.
one OFDM block.
Eric Pierre Simon et. all[3]proposed joint carrier fre-
The blind channel estimation is carried out by evaluating
quency offset(CFO) and channel estimation for OFDM system
the statistical information of the channel and particular prop-
over the fast time varying frequency selective channel using
erties of the transmitted signals. This blind channel estimation
Expectation-Maximization (EM) Algorithm. This algorithm
has no overhead loss and it is only suitable for slowly time-
jointly estimates path channel gain and CFO in the presence of
varying channels. But in training based channel estimation,
high mobility. The time varying channel gain in each OFDM
training symbols or pilot tones that are known to the receiver,
symbol is approximated by a basis expansion model (BEM)
are multiplexed along with the data stream for channel esti-
representation. This algorithm is better efficient than existing
mation.
algorithm for fast time varying channels.

IJSER © 2013
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 4, Issue 5, May-2013 1854
ISSN 2229-5518

A Channel estimator using neural network is presented by A block fading channel is a channel which is constant over
A. Omri et.all [4] for Long Term evolution (LTE) uplink. This a few OFDM symbols. In this channel, pilots are transmitted
method uses knowledge of pilot channel properties to esti- on all sub carriers in periodic intervals of OFDM blocks. This
mate the unknown channel response at non-pilot subcarriers. type of pilot arrangement is shown in Fig. 3(b), is called as
This type of estimator learns to adapt to the channel variations block type arrangement.
and then it estimates the channel frequency response. This
method is less complex and high quality than conventional
methods such as least Square (LS), Minimum Mean Square
Error (MMSE). Also this method has more mobility.
Ahamed Gomaa et. all.[5] presented a novel ap-
proach based on compressive sensing(CS) theory to estimate
and mitigate synchronous narrow band interference(NBI) in
MIMO-OFDM system. Compressive sensing theory showed
how to reconstruct a sparse vector from a noisy environment.
In this approach, NBI is first estimated and cancelled before
Figure .3(a) Comb Type Pilot
channel estimation. This technique has less performance loss
due to the channel estimation errors. This approach is studied
for time varying & frequency selective channels. This NBI es-
timation can be applied to both zero padded- OFDM and Cy-
clic Prefix –OFDM systems.
A semi blind algorithm is presented by Feng Wan
et.all [6] for the estimation of sparse MIMO-OFDM system. In
this approach , a second order statistics of signal received

IJSER
through sparse channel is expressed in terms of most signifi-
cant taps(MSTs) of the sparse channel. Blind constraint for the
channel is derived from the MST position and this constraint is
then combined with the training based least square method to
develop a semi blind algorithm. Simulation results of this Figure .3(b) Block type Pilot
method shows that performance of this method is better than
other blind algorithms. This method has also been extended The MIMO-OFDM system performance is evaluated by
for sparse channels in the up-sampling domain for MIMO- means of the plot of Mean Square Error (MSE) and Bit Error
OFDM system. Rate (BER). Block-type and comb type pilot based channel
Nima Sarmadi et.all[7] proposed a new blind channel estimation using LS and MMSE algorithms [9] are used to
estimation algorithm for MIMO-OFDM system.This method model Rayleigh fading channel of MIMO-OFDM systemand
uses specific properties of orthogonal space time block codes Mean square error is estimated.The MMSE channel estima-
to estimate the finite impulse response channel parameters in tion has low Mean Square Error than LS channel estimation
time domain.A semi definite relaxation technique is used for algorithm in this approach.
estimation in this approach.This technique has less complexity
and good performance than existing blind channel estimation 4.2 Optimization Techniques
techniques. Various optimization techniques are used [10],[11],[12] to op-
timize the placement of pilots,power and LS & MMSE algo-
4.1 Training Based Channel Estimation Techniques rithms.
The training-based channel estimation can be performed by
either block type pilots or comb type pilots.Least Square(LS), Muhammet Nuri Seyman et.all.[10] proposed particle
Minimum Mean Square Error estimation techniques and swarm optimization (PSO) to optimize the placement and
their performance is discussed in [8],[9]. In block type pilot power of comb-type pilot tones used for LS channel estima-
estimation, pilot tones are inserted into each OFDM symbol tion in MIMO-OFDM system.Mean square error is used as an
with a specific period of frequency bins. This type of channel objective function of PSO in this technique.This approach has
estimation is very much suitable where the changes even in less complexity and better placement of pilot tones with min-
one OFDM block. imum error.

A fast fading channel has channel impulse response K.Vidhya et.all [11] proposed evolutionary programming
changes rapidly within symbol duration and the channel technique to optimize LS & MMSE algorithms for better re-
changes between adjacent OFDM symbols. In this fast fading, sults.In this approach ,evoloutionary programming method of
the pilots are transmitted at all times but with an even spacing mutation & crossover operations are applied to the existing
on the sub carriers, representing a comb type pilot placement algorithms and best channel estimation matrix is de-
which is shown in Fig. 3(a) .The channel estimates from the rived.Experimental results of this approach shows that this
pilot sub carriers are interpolated to estimate the channel at enhanced channel estimation outperforms the LS & MMSE
the data subcarriers. estimation techniques.

IJSER © 2013
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 4, Issue 5, May-2013 1855
ISSN 2229-5518

Muhammet Nuri Seyman et.all.[12] proposed differential [11] K.Vidhya K.R.Shankar Kumar “ Enhanced Channel Estimation Tech-
evolution algorithm to optimize pilot tones in MIMO-OFDM nique for MIMO-OFDM Systems with the aid of EP Techniques” in
system.simulations of this approach shows that the perfor- European Journal of Scientific Research .Vol.67,No.1,2011 pp. 140-
mance of LS algorithm was increases by optimizing pilot tones 156.
with the help of diiferential evolution algorithm. [12] Muhammet Nuri Seyman,,Necmi taspinar “ Optimization of Pilot
tones using diiferntial evolution algorithm in MIMO-OFDM Sys-
tems” in Turk Journal of Elec Engg & Comp SCi. Vol.20. No.1, .2012,
5 CONCLUSION
pp. 15-23.
In this paper,the basic concepts of Orthogonal Frequency
Division Multiplexing (OFDM), Multiple Input Multiple Out- [13] http://www.wikipedia.org/MIMO
put(MIMO) systems are discussed. The various channel esti-
mation techniques such as training based,blind channel,semi-
blind channel based algoritms and their performance are also
discussed.Also different optimization techniques such as par-
ticle swarm optimization,evolutionary programming is
Reviewed to optimize LS & MMSE algorithms.

REFERENCES
[1] Sivash M. Alamouti “ A Simple Transmit Diversity Technique for
Wireless Communications ”in IEEE journal on Select Areas in Com-
munications,Vol.16. No.8, Oct.1998 pp. 1451-1458.
[2] Eric Pierre Simon ,Laurent Ros, Hussein Hijazi, Jin Fang, Davy paul

IJSER
Gaillot, Marion Berbineau “ Joint Carrier Frequency Offset and Fast
Time varying Channel estimation for MIMO-OFDM Systems” in
IEEE Transaction on Vechicular technology, Vol.60. No.3, Mar.2011,
pp. 955-965.
[3] Eric Pierre Simon ,Laurent Ros, Hussein Hijazi, Mounir Ghogho, “
Joint Carrier Frequency Offset and Channel Estimation for OFDM
Systems via the EM Algorithm in the Presence of Very High Mobili-
ty” in IEEE Transaction on Signal Processing, Vol.6.No.2, Feb.2012,
pp.754-765.
[4] A.Omri , R.Bouallegue, R.Hamila and M.Hasna “ Channel Estimation
for LTE Uplink System by perceptron Neural Network” in Interna-
tional Journal of Wireless & Mobile Networks, Vol.2, No.3, Aug .2010,
pp. 155-165.
[5] Ahamed Gomaa , Naofal Al-Dhahir “ A Sparsity-Aware Approach
for NBI Estimation in MIMO-OFDM” in IEEE Transaction on Wire-
less communications. Vol.10. No.6. June 2011. pp. 1854- 1862.
[6] Feng Wan,Wei-Ping Zhu, M.N.S.Swamy , “ Semiblind Sparse Chan-
nel Estimation for MIMO-OFDM Systems”, in IEEE Transaction on
Vechicular Technology,Vol..60, No.6 July 2011.
[7] Nima sarmadi,Shahram Shahbazpanahi,Alex B.gershman “Blind
channel Estimation in Orthogononally coded MIMO-OFDM
Systems: A Semi definite Relaxation Approach” IEEE Transcation on
Signal Processing,vol.57,No.6,June 2009,pp.2354-2364.
[8] Kala Praveen Bagadi,Susmita Das,”MIMO-OFDM Channel Estima
tion Using Pilot Carriers”in International Journal of Computer
Applications,Vol.2,No.3,May2010,pp.81-86.
[9] R.S.Ganesh,J.Jaya Kumari,Akhila I.P,”Channel Estimation Analysis
in MIMO-OFDM Wireless System” in Proceedings of 2011 IEEE
International Conference on Signal Processing, Communication,
Computing and Networking Technologies (ICSCCN 2011),pp.467-
471
[10] Muhammet Nuri Seyman,,Necmi taspinar “ Partcile Swarm Optimi-
zation for Pilot Tones design in MIMO-OFDM Systems” in Eurasip
Journal on Advances in Signal Processing 2011.

IJSER © 2013
http://www.ijser.org