Research paper (channel_estimation)

CHANNEL ESTIMATION IN MIMO OFDM USING DISCRETE FOURIER
TRANSFORM WITH DENOISE MODEL BASED LEAST SQUARE CHANNEL
ESTIMATOR
Abstract
In Multiple-input and Multiple-Output (MIMO) system Orthogonal frequency
division multiplexing (OFDM) system, is utilized for effective operation of multipath
communication. The channel estimation (CE) is utilized for demanding channel condition
with time-varying characters. The existing channel estimation techniques are incredibly
complex. To resolve this complexity in CE methodologies, this paper developed as Improved
Channel Estimation Algorithm integrated with LS DFT (ICEA - DA) is proposed. On
transmitter block, QPSK scheme is utilized for modulating input signal with concept of pulse
shaping and Least square approach. The LS Estimation technique needs the channel to be
consistent all through the period of estimation. Minimum Mean Square Error (MMSE)
requires knowledge of channel statistics (KCS). Performance measure of MSE of channel
estimator fusing DFT based procedure obtains best results as contrasted with consolidation of
LS based strategies. In transmitter side, information from transmitter to receiver is
transmitted by mode of AWGN (Additive White Gaussian Noise) channel. This operation is
performed inversely in receiver block and cost function is lessened with use of improved
differential evolution technique (IDE). In this paper, channel estimation strategy is proposed
in MIMO OFDM utilizing LS with DFT.DFT join with LS gives higher estimation precision
as well as limiting MSE and BER. Experimental analysis of proposed approach with existing
technique stated that ICEA - DA provides significant performance in terms of SER (Symbol
error rate), BER (Bit error rate), CC (Channel Capacity) and PSNR (Peak Signal to Noise
Ratio).
Key words: MIMO, OFDM, Least Square Estimator, Channel Estimation, Minimum
Mean Square Error, Discrete Fourier Transform
Introduction
Enabling of 5G technology is recognized by MIMO in mobile communications [1].
Performance gain in extensive MIMO can be attained, and can be reliable for the exposure of
channel state information (CSI) is important [2]. Number of downlink CSI in acquiring CSI
conventionally since they can use the large scale array antenna in base station even in
downlink massive MIMO [3] then, array size of base station is proportional, the channel

matrices are not be able to computation requires extremely maximized training for downlink
overheads since it is direct estimation.
MIMO transmission will be the major part of wireless communication in forth coming
days in wireless technology. This type of system has minimum difficulty in executing OFDM
and it simple while comparing, in case of non-orthogonal multi-carriers MIMO structure of
receiver has become more complicated since it has to handle with various interference
measurements, i.e. inter-symbol- interference (ISI), inter-carrier- interference (ICI) and inter-
antenna interference (IAI). Those reference signals that are received has been influenced by
various measurements of interference (known as pilots), this makes the estimation of the
channel (CE) complicated and consecutive equalization (EQ) of the received signal (Rx),
mostly, the channel is dispersed in multiple types of a broadband communication.
Channel Estimation (CE) utilized Channel State Information (CSI) for transmission and
reception of wireless signal. In wireless communication medium, CSI leads to scattering of
wave, fading and power decay with respect to transmission distance. The concept of CE is
evolved based on the objective of correlation of signal at center point integrated with
suppression of MAI. This signal quality impacts on receiver BER performance [7]. In CE two
schemes are employed for estimation of signal such as (i) pilot assisted scheme where,
training symbols are allotted for bandwidth and (ii) Blind Approach with utilization of
statistical features. In receiver end, CE perform insertion of pilot signal within the transmitted
signal where, pilot signal leads to signal overhead, this can be reduced by minimal number of
pilot symbols.
At present, for uplink MU-MIMO communication Innumerable MIMO detectors are
utilized. Also, optimum Maximum Likelihood detector and low-complexity environment are
developed. In [4] it is studied that the improved channel estimation of CP information is the
design to utilize through kalman filters. Though this techniques can be employed only for
waveforms of orthogonal. So this paper took the advantage of localization of symbol-time in
waveforms that are orthogonal just to enhance the operation of CE through reusing the
information of pilots from CP.
It can also be referred by localizing the pilots energy up to the terminated point of the block
properly, it need to possess maximum energy while it is copied to CP. Identification of
symbols in data of non-orthogonal multi-carriers acquires various CE as well as CQ modules
that are independent to the usual practice. The earlier works on CE [5], it has to be

calculating LS then LMMSE estimators that assume ISI and ICI manages to additional noise
of correlation with the white noise of non-orthogonal waveform. Even though the techniques
that are non iterative has been discussed in [5] that can acquire implementation of CE with
higher errors and SNR that are suboptimal. It has no solution for channel estimation literally
and can be applied to interference of MIMO directly with limited interference and non
orthogonal waveforms.
In this paper, proposed a channel estimation method with integration of differential
evolution technique and LS with DFT. The proposed ICHE-DA included in central part of
OFDM - MIMO block. We consider two-user interference channel for MIMO-OFDM
systems. The main idea is to take into account the difference between the intrusive pilots and
the intrusive data signal in terms of optimized concept. We demonstrate the basic idea of user
based transmitter and receiver using least square estimation with discrete Fourier method
(LS-DFT). Further, we focused on interference from frequency-selective channels. By doing
so, the main struggle is to evaluate the accuracy of covariance matrices, which may need
proper information of interference parameter. Moreover, one of the contributions is to
analyze the firmness of the spectrum by coinciding with DFT series. Results illustrated that
proposed scheme exhibits significant performance rather than existing technique in terms of
BER, PSNR, CC and MMSE.
Literature Survey
Farnoosh Talaei et al.,(2019)introduced the MIMO-OFDM for hybrid channel
estimator in which the sparse nature for channel that are angular and with compressed sensing
(CS) tools on basis of Wave systems are noted. Limited angular resolution has been solved by
angular channel with continuous framework by action as the discretization has been also
resolved by angular support of sparse channel designs that are utilized in CS on basis of
channel estimation. Minimum construction error and optimal performance for proposed
channel estimator is calculated by performance analysis [20].Wei Ji et al.,(2018) discussed
the novel scheme for channel estimation in downlink channels by CSMBCS technique in
channel estimation of downlink. The usual sparsity and cluster design has been employed
where the user is not known, just to evaluate the scenario of various subcarriers of downlink
channels. The local initial clusters has been constructed using architecture of Bayesian has
been described with the property of sparsity and local beta process (LBP) [21].

Shahab Ehsanfar et al.,(2018) determined parallel-interference-cancellation (PIC)
technique on basis of MIMO LMMSE for shared estimation of channel as well as leveling of
non-symmetrical waveforms. In contrast to the basic practice, by appropriately restricting the
pilots in time area, we likewise utilize pilots' data from CP. Through execution results, They
demonstrated that utilizing CP data of pilots for increases up to 2.4 dB preferable edge error
rate execution over OFDM signal [22]. Javad Mirzaei et al., (2019) proposed semi-daze,
time-space, estimation of channel procedure for frequency-selective enormous MIMO
frameworks. Arrangement relies upon subspace spreaded by signed eigenvectors obtained
from signed covariance matrix. Critically, the receiver examines the reception rate based on
time-space estimation accessed for an exact matrix estimation. Estimation technique does not
need symmetry between preparation images of clients in all cells [23].
Yuan Huang et al., (2018) anticipated the NAMP calculation just as assessment model
which depends on LTE-Advanced remote channel model. At first, NAMP needn't bother with
the sparsity level priori-information. At that point, decide the fixed advance size that
improves the sign reproduction effectiveness. At last, a Singular Entropy request assurance
component which endeavors to stay away from the less pertinent iotas that can be presented.
It accomplishes increasingly stable execution particularly in low SNR channel condition, and
after that the computational multifaceted nature is decreased like versatile SAMP calculation
[24].
The estimation of channel covariance matrix of the received sampled signals in
frequency-domain is proposed by Ngo et al., [25]. In an asymptotically massive MIMO
systems channel vectors possess orthogonal property, where channel vector and eigen vector
are proportional to each other. The covariance matrix estimation by limited number of signal
samples produces error and assumption of channels are perfectly orthogonal are the two
limitations of the system. the condition assumed in this method is both the samples and
antennas numbers are asymptotically large.
Zarifi et al., [26] proceeded for CDMA system having unknown WSS noise. For the
estimation of noise subspace in data samples provided, generalized correlation
decomposition-based blind channel estimation noise covariance matrix is estimated by the
centro- Hermitian property. Lamare et al.,[27], considered multipath channel for STBC-
CDMA system. STBC is implemented by exploitation of spreading codes having unique
structure and developed computationally efficient based channel estimation algorithm.

The angular resolution is improved by utilizing a redundancy samples containing
substantially more uniform DSFT premises the antenna numbers is proposed by Lee et al.,
[28]. DSFT number premises, is equal to angular spectrum samples at increased rate which
increases both length of insufficient channel vector need to be recuperated and quantity of
nonzero components are improved. Along these lines, more channel estimations (longer
preparing groupings) should be acquired for extraordinary recuperation of the signal.
Minimum Mean Square Error (MMSE) scheme with second order statistics is
proposed by Li et al., [29]. Limitations of this method are the channel estimation error. Joint
estimation of channel length as well as OFDM system impulse response is proposed by
Charles et.al [30], making use of the balanced achievement between information specification
based on Kullback-Leibler divergence and noise rejection with accurate channel description.
Hence, data specification is not suitable for practical channel length estimation techniques
because of prohibitive complexity. Estimation of channel is performed in a recursive manner
allowing optimal channel length establishment with considerable increase in cost and
increased performance and robustness.
Proposed Methodology
This paper presented a improved channel estimation algorithm (ICEA-DA) integrated
with differential approach. Also, this paper uses LM.DFT approach for channel estimation in
wireless communication medium. The proposed (ICEA-DA) placed in the central part of
MIMO framework. Generally, MIMO system comprises of multiple transmitter and receiver
antenna for parallel data transmission. The proposed model is examined for 2 transmitter and
4 receiver antenna. In transmitter side, Quadrature Phase-Shift Keying modulation (QPSK)is
utilized for input modulation. Then,ISI is mitigated through utilization of PSA, where PSA
involved in transmission of waveform pulses and effective utilization of available bandwidth.
In case,signal bandwidth more effective rather than channel bandwidth, it leads to signal
distortion. This distortion are observed in ISI environment. To control ISI in wireless
communication pulse shaping filter (PSF)is utilized for analysis. At each transmitter side,
symbol mapping and IFFT operations are performed. Subsequently, it leads to transmission
of information over Multipath channel and includes AWGN noise in receiver side. At the
receiver, this operation is performed inversely with utilization of ICEA-DA. With utilization
of LMMSE and DFT semi-blind review is performed. In figure 1 detailed overall flow of
constructed MIMO system with ICEA-DA is presented.

Figure 2: Proposed block diagram on OFDM for transmitting and receiving process
Initially,for the incoming signal information are transmitted through inclusion of pilot
sequences and modulation scheme such as QPSK scheme. The QPSK scheme includes
consists of advantages and noise factor. Another advantage of this scheme is higher
information transmission rate rather than other modulation scheme with minimal bandwidth.
The signal shift consists of ‘4’ signal states with transmission of information through QPSK
scheme as 2bits/symbol. Through inclusion of coding scheme every signal in system varies in
1-bit. In below equation (1) signal transmitted in sine and cosine form is presented as follows:
    4,3,2,1,
4
122cos
2






 nntf
D
S
tg c
s
e
n


(1)
Here, signifies the signal based on the time, represents the symbol energy;
symbol duration is denoted as , whereas, is represented as baseband of signal.
 tgn eS
sD cf
Data S(n)
input
Pilots
insertion and
space time
encoder
s/p
(negati
ve)
s/p
(positiv
e)
ICEA-DA
DFT with
FFT
IDFT
P/
S
P/
S
upconver
ter
upconver
ter
transmitter
receiverDown
converter
Remove
CP
DFT
Space
time
decoder
Decod
ed data
s*(n)
output
Channel
estimator

In modulator of MIMO scheme 4 phases of signal are generated which consists of 2D
signal space with unit function  t1 and  t2 denoted as following equation (2) and (3)
  Tttf
D
t c
s
 0,2cos
2
1 
(2)
  Tttf
D
t c
s
 0,2sin
2
2 
(3)
From above equation (2) and equation (3) and utilized for measurement of in-
phase and quadrature component. In terms, OFDM signal consists of 4 signal constellation
based on signal-space four points such as defined in equation (4):









2
,
2
ee SS
(4)
In this, the signal factor 1/2 signifies the total power of the system which has been
partitioned equally between inclusion of 2 carriers. In receiver side, the symbol is
demodulated with elimination of carrier-phase factor, the phase 2 successive received bots
are ascertaining with respect to input data.
Let  tpt indicates the transmit-side PSF for each symbol and  tpr signifies the receive
antenna side matched the filter. The composite channel symbolized as  tH . The  TR II , the
time domain signal received at the Ri th
antenna as follows:
   

T
T
N
i
iTiTiRiTiR Nxnhnh
1
,, (5)
Here, in equation (5) represents denotes noise placed in uncorrelated spatio temporal
region, represents presence of noise in transmit time-domain signal at the ith antenna.
After removal of ISI, with utilization of IFFT ISI is eliminated also it uses symbol mapping
technique at each transmitter side. In frequency domain components fourier analysis is
performed for signal transmutes in time or spatial domain factors. At receiver end, IFFT is
implemented for improving flexibility, execution speed and precision. With utilization of
IFFT operation various complex and data points are transmuted for same number of points in
 t1  t2
N
iTx

time-domain. The IFFT involved in execution of N-point receiver operation in equation (6)
IFFT is presented mathematically as follows:
    1,.....,1,0,
1 1
1
 



NnekU
N
nu
N
n
N
knj
(6)
Where represented data frame transform size or several points in the signal.
denotes the FFT frequency output at kth
point where k=0,1,…….., N-1.Subsequently, FFT
analysis is utilized for transmute of signal in time domain at receiver block. In equation (7)
mathematical formulation for FFT is presented as follows:
   




1
1
2
1,.....,1,0,
N
n
N
knj
NkenukU

(7)
Additionally, in multipath channel environment signal are transmitted with addition of
AWGN noise characteristics. In this paper, AWGN is utilized for its higher bandwidth rather
than message signal. The system noise n represents complex Gaussian distribute random
stationary point with mean 0 and formulate vector with identical components with message
define in equation (8) as below:
    1,01,0 NiNnn p  (8)
Where, signifies the noise power and represented message signal of same
length which comprises of normal/Gaussian random variables.
Firmness of spectrum:
The basic frequency that is used for the approximation of the given signal is f0 =
1/T0, where T0 is the period of the signal. For time-discrete signals this corresponds to a
frequency of 1/N. Since all other frequencies are multiples of this basic frequency, the
number of samples N defines the resolution of the spectrum. The maximum possible signal
frequency represented in X[k] is at k = ⌊ N/2⌋ . Assuming the original signal is an audio
signal sampled at 44.1 kHz, then it contains a maximum frequency of 44.1/2=22.05 kHz
according to Shannon’s sampling theorem. Taking N = 1024 samples, for example, the
coefficient at k = N/2 corresponds to a frequency of 22.05 kHz and the coefficient at k = 1
corresponds to a frequency of 22.05/512 ≈ 43.1 Hz. If we would sample another piece of
audio signal with only 1000 samples (at same sampling rate), the DFT coefficients would
N  kU
pn   1,0N

represent other frequencies (e.g. k = 1 would correspond to 22.05/500 = 44.1 Hz). The
conclusion is that spectra can only be compared when the number of samples, from which the
DFT spectrum is computed, is the same. Another effect regarding the resolution has been
demonstrated by the second numerical example. Since the DFT spectrum is sampled, i.e.
there is only information about certain discrete frequencies, it cannot not reflect the true
contributions of the single frequencies of an arbitrary signal x[n], even this signal is periodic.
Receiver structure:
Using LS-DFT with superimposed pilots, receiver with channel estimation and MIMO
detection is proposed. Without generality loss, it is considered that there is pilot for every
subcarrier of every block of frame , i.e., ΔNF = ΔNT = 1, leading to N(TS) = NNT and pilot
multiplicity or redundancy of
NR = N (TS)/NG
At receiver side, due to phase difference belongings of received signal causes multipath
fading. It happened due to the received signals travelled in different distances and paths. The
Rayleigh distributions are used which reaches at a receiver by more than one path. Rayleigh
distribution is most normally used for fast fading case. The pilot assignment is performed by
reuse the pilot sequence for own pilot request from multiple request. The proposed
optimization algorithm easily optimizes the time varying channel metrics. The proposed pilot
assignment method can offer a powerful way to compute the requested user in the own cell.
The fitness function (FF) is an important for optimization algorithm. In this paper, the pilot
assignment is performed by reuse the pilot sequence for own pilot request from multiple
request. The utilized optimization algorithm easily optimizes the time varying channel
metrics.
The proposed pilot assignment method can offer a powerful way to compute the
requested user in the own cell. The fitness function (FF) is an important for optimized
concept in receiver side.
Improved Channel Estimation Algorithm using differential evolution
Here, proposed ICEA - DA is based on the consideration of cost function for obtaining
solution for parallel search optimization and continuous process. The advantage of proposed
ICEA - DA is improved convergence through utilization of control parameters. The proposed

ICEA - DA is represented in canonical form with effective handling of continuous variables
with optimization of integer variables. The proposed ICEA - DA utilized for employing
MMSE estimation as objective form. In the ICEA - DA approach MMSE is considered as
objective function with inclusion of several steps such as initialization, objective function and
mutation. Initially, populace NP is considered as population size for initialization of
parameters. Apart from those parameters, for constructed function random individuals are
provided with lower-upper bounds represented in equation (9)
   Djvvrandvv jjjj
goi ,....,3,2,1;.1,0 minmaxmin.  (9)
In above equation (9) represented individual index, denoted lower-upper bounds
with comprises of D parameters those are generated uniformly with inclusion of equation
(10) as follows:
   DD
vvVvvV max
1
maxmaxmin
1
minmin ,......,and,......,  (10)
Next, indicates the random generator between the range of 0 to 1.
For evaluating fitness function ICEA - DA is utilized for estimation of MMSE of OFDM
system. It is represented in equation (11) as below:
   LS
ffHfHfHHf
MMSE
f hISNRRRH
1
/

  (11)
Subsequently, based on the above equation (11) mutation is performed in MIMO OFDM
system. This differential evolution is genetic methodology to withstand variation in
population based on generation. Based on constructed ICEA - DA new parameters are
summed between two parameters with estimation of weighted difference factor based on
temporary, trial or third population. For the generated vector scheme trail population G is
represented in equation (12):
 GVGVVW rrGrGi .. 32.11,   (12)
Where indicates a target vector   321321 ,,.,......3,2,1,,, rrrNPrrri  of the signal with
MIMO OFDM system. Moreover, in proposed approach indices arbitrarily point indices are
presented with scaling factor are with differential variation of [0,1]. After that,
j iv
 1,0rand
V


termination-state of MMSE is utilized for identification of effective solution for MMSE
through maximal number of generation.
RESULTS AND DISCUSSIONS
This section provides evaluation and comparison of proposed ICEA - DA algorithm
performance with other system are presented. The performance metrics considered for
analysis are BER, MSE and MMSE utilized for evaluation of proposed approach system
performance. The proposed ICEA - DA is network measurement is observed by varying SNR
of system with respect to number of sender and receiver antenna. The proposed ICEA with
LS and DFT is executed in MATLAB platform with machine configuration of CPY Speed
3.20 GHZ, RAM 4 GB and Windows 7 OS.
Performance analysis
The simulation parameters used in the proposed method moreover, their values are
shown in Table 1.
Table 1: Simulation parameters
No of FFT points 1085
Length of Cyclic Prefix 648
Total no of Subcarriers 5385
Total no of Symbol 289
Pilot Arrangement Recursive type
Pilot Constellation QPSK
Data Constellation QPSK
Bandwidth 9MHZ
SNR Range 12 db
Number of Iterations 300
Here, the proposed ICEA - DA with LS FFT is utilized for analyzing performance of
system. The metrics considered for analysis are expressed as follows:
 Bit error rate (BER)
In characterizing data channel performance, it is used as important parameter. At
remote end, errors in key parameters are appeared while transmitting data from one to
another point in wireless link/radio or wired telecommunication link.

 
 bN
b
tT
eN
BER 
Where, Number of bits in error, and
total number of transmitted bits.
 Mean square error
It is average squared difference between estimated values and actual values. It
is risk function corresponds to expected value of squared error loss.
Table 1: Mean square error (MSE) and Bit Error Rate (BER) of hybrid LS with DFT
SNR BER MSE
0 0.088385 0.015678
3 0.034779 0.007881
6 0.010104 0.003916
9 0.002036 0.001961
12 0.000367 0.000987
15 5.47E-05 0.000494
The below Fig. 3 shows the simulation as well as analytical results comparison which obtain
the MSE as well as BER of equalizer output by DFT-based then Wiener channel estimators.
The analytical as well as semi-analytical approaches for producing MSE related equalizer
output from simulation is obtained. Discrepancy slightly observed with high SNR (> 15 dB)
designed for DFT-based channel estimator.
 eNb
 bN tT

Figure 3: graphical representation of MSE and BER for proposed method
Table 2: Mean square error (MSE) of Minimum Mean Square and Least Square estimators
SNR MSE of LS MSE of MMSE
0 0.115917 0.028968
3 0.065559 0.020203
6 0.031985 0.011584
9 0.015689 0.006628
12 0.007218 0.003304
15 0.004017 0.001777

Figure 4: MSE values of LS, MMSE and LS-DFT method
Table 3: Bit error rate (BER) of minimum mean square and least square estimators
SNR BER of LS BER of MMSE
0 0.089109 0.075906
3 0.073078 0.054953
6 0.066734 0.043031
9 0.064219 0.037188
12 0.065094 0.032375
15 0.069031 0.030234
Figure 5: BER values of LS, MMSE and LS-DFT method

Figure 4 and 5 illustrates LS and MMSE channel estimation methods performance of
MSE and BER. The experimental results represents that the estimator gains of MMSE which
is more than 4 dB larger than LS estimator through all values of SNR. Because of MMSE
ability which is remove the noise outside. Additionally, the value of SNR get high then the
MMSE estimator of MSE becomes low when compared to LS estimation.
Table 4: MSE and BER values of DFT, Harmonic retrieval and LS-DFT method
parameters Harmonic retrieval
method
DFT LS with DFT
MSE 0.1689 0.004017 0.4310
BER 0.08654 0.069031 0.4198
Number of complex computation of conventional SLM OFDM when number of carriers
is N = 2n
, is given by
Nmul =2n-1
nU
Number of complex additions for type1 system is given as:
Complexity = Nlog2N + 3(U-1)N
Computational complexity of type-2 method is efficiently minimized and it is given as
Complexity = 2Nlog2N + 3(U-2)N
Table 5: Conventional complexity of DFT, Harmonic retrieval and LS-DFT method
Number of
blocks
Complex
computation
Harmonic retrieval
method
DFT LS with
DFT
4 Rotation factors 427930(23.3%) 855900 (46.6%) 19.3%
Multiplication 126362 (52.9%) 138650(48.2%) 39.6%
Addition 1104500(28.1%) 1850400(47.1%) 21.4%
8 Rotation factors 102040 (20.4%) 267893 (40.1%) 16.5%
Multiplication 299008(46.1%) 303104(51.4%) 40%
Addition 237568(20.2%) 245760(45.5%) 17.1%

Inference
The performance of proposed approach is comparatively examined with existing
technique such as QAM, BPSK and without modulation scheme. The parameter considered
for analysis are LS, MMSE, BER and PSNR. The comparison details are presented in below
section.
Figures 3, 4 and 5 represent appropriateness of channel estimator in OFDM. BER
behavior of LSDFT is completely different when observe OFDM in which its value tends to
zero when SNR increases. There are two reasons. (i) LS-DFT is utilized as wiener filter when
residual interference remains constant. (ii) with selectivity, number of holes in frequency
response increases. ZF equalizer coefficients Fm = 1/Hˆm are used for estimation errors, for
high frequency equalization stage causes more errors in detected symbols.
In figure 6 performance of proposed method is presented based on the consideration of
various modulation scheme in to consideration with inclusion of PSNR measurement. In the
condition, PSNR value is observed as proposed ICEA - DA approach with LS DFT provides
PSNR of 42, while other scheme such as QAM and BPSK provides modulation value of 9
and 23 respectively. In case of without modulation PSNR value is measured as 28. It is
concluded that proposed scheme exhibits significant performance.
Figure 6: Comparison of PSNR for proposed with other techniques

In figure 7 performance of proposed approach is comparatively examined with other
existing technique such as QAM and BPSK scheme, For other methods, QAM, BPSK and
without any modulation scheme BER exists between 103
104
to SNR gain of 1dB to 11dB. In
case of proposed method BER exists between 101 to 102 with SNR of 10dB and 11dB with
significant performance. Hence it is concluded that proposed scheme exhibits significant
performance rather than other existing techniques.
Figure 7; Comparison of SER for proposed with existing technique
Figure 8: Comparison of system channel capacity

In figure 8 channel capacity is presented with variation in number of receiver. For NR 2
CC lies between 100 to 10 with SNR gain of 1dB to 11 dB gain. When NR is 2 CC is below
101 from this it is observed that NR increases with increase in CC. Consequently, it is
concluded that proposed approach exhibits significant performance.
Conclusion
CE is considered as promising technology with incorporation of MIMO. This paper
proposed a ICEA- DA approach for improving channel capacity estimation with inclusion of
LS with DFT concept in existing OFDM technique. As a result, the results showed that the
least square is about SNR of 0.68V and MMSE calculation around 0.089V is obtained. The
channel estimation errors effect on MSE is analyzed equalizer output. MMSE estimation
execution is better as far as MSE. On analysis of channel capacity it is observed that SNR
gain increases with increase in number of antenna. Experimental results inferred that the
proposed ICEA - DA provides significant performance compared with existing technique. In
future, this work can be extended with consideration of various number of transmitter and
receiver for specific applications.
Conflict of Interest:
Authors declare that ‘no conflict of interest’ in preparing this research article
Funding:
This Research Received no specific grant from any funding agency in public, commercial or
not-for-profit sectors.
Reference
[1] E. Hossain and M. Hasan (2015), “5G cellular: key enabling technologies and
research challenges,” IEEE Instrumentation & Measurement Magazine, 18(3):11–21.
[2] E. G. Larsson, O. Edfors, F. Tufvesson, and T. L. Marzetta (2014), “Massive MIMO
for next generation wireless systems,” IEEE communications magazine, 52( 2):186–
195.
[3] Z. Jiang, A. F. Molisch, G. Caire, and Z. Niu (2015), “Achievable rates of FDD
massive MIMO systems with spatial channel correlation,” IEEE Transactions on
Wireless Communications, 14(5): 2868– 2882.

[4] A. K. Kohli and D. S. Kapoor (2016), “Adaptive Filtering Techniques Using Cyclic
Prefix in OFDM Systems for Multipath Fading Channel Prediction,” Circuits Syst.
Signal Process., 35(10): 3595–3618.
[5] S. Ehsanfar, M. Matthe, D. Zhang, and G. Fettweis (2016), “A Study of Pilot- ´ Aided
Channel Estimation in MIMO-GFDM Systems,” in Proc. of the ITG/IEEE Workshop
on Smart Antennas (WSA’16).
[6] R. W. Heath, N. Gonzalez-Prelcic, S. Rangan, W. Roh, and A. M. Sayeed (2016), “An
overview of signal processing techniques for millimeter wave MIMO systems,” IEEE
Journal of Selected Topics in Signal Processing, 10(3):436–453.
[7] J. Lee, G.-T. Gil, and Y. H. Lee (2016), “Channel estimation via orthogonal matching
pursuit for hybrid MIMO systems in millimeter wave communications,” IEEE
Transactions on Communications, 64(6):2370–2386.
[8] A. Alkhateeb, O. E. Ayach, G. Leus, and R. W. Heath, “Channel estimation and
hybrid precoding for millimeter wave cellular systems,” IEEE Journal of Selected
Topics in Signal Processing, vol. 8, no. 5, pp. 831–846, Oct. 2014.
[9] D. Ramasamy, S. Venkateswaran, and U. Madhow, “Compressive tracking with
1000-element arrays: A framework for multi-Gbps mm wave cellular downlinks,” in
Proc. IEEE Annual Allerton Conference on Communication Control and Computing,
Monticello, IL, USA, Oct. 2012, pp. 690–697.
[10] D. E. Berraki, S. M. D. Armour, and A. R. Nix, “Application of compressive
sensing in sparse spatial channel recovery for beamforming in mmwave outdoor
systems,” in Proc. IEEE wireless communication network conference, Istanbul,
Turkey, Apr. 2014, pp. 887–892.
[11] A. Alkhateeb, G. Leus, and R. W. Heath, “Compressed-sensing based multi-user
millimeter wave systems: How many measurements are needed?” in Proc. IEEE
International Conference on Acoustics, Speech and Signal Processing (ICASSP),
Brisbane, QLD, Australia, Apr. 2015, pp. 2909–2913.
[12] R. Mendez-Rial, C. Rusu, A. Alkhateeb, N. Gonzalez-Prelcic, and R. W. Heath,
“Channel estimation and hybrid combining for mmwave: Phase shifters or switches?”
in Proc. IEEE Information Theory and Applications, San Diego, CA, USA, Feb. 2015,
pp. 90–97.
[13] Y. Li, L. J. Cimini, Jr., and N. R. Sollenberger, “Robust channel estimation for
OFDM systems with rapid dispersive fading channels,” IEEE Trans. Comm., vol. 46,
no. 7, pp. 902-915, July 2004.

[14] Edfors, M. Sandell, J. J. van der Beek, S. K. Wilson, and P. O. Borgesson, “OFDM
channel estimation by singular value decomposition,” IEEE Trans. Comm., vol. 46,
no. 7, pp. 931-939, July 2015.
[15] S. Haykin, Adaptive Filter Theory (4-th Edition), Prentice Hall, 2002
[16] Y. Zhao and A. Huang, “A novel channel estimation method for OFDM mobile
communication systems based on pilot signals and transformdomain processing,”
Proc. 2015, pp. 2089-2094.
[17] B. Yang, K. B. Letaief, R. S. Cheng, and Z. Cao, “Windowed DFT based pilot-
symbol-aided channel estimation for OFDM systems in multipath fading channels,”
Proc. VTC’00-Spring, vol. 2, pp.1480-1484.
[18] W. C. Jakes, Microwave Mobile Communications, John Wiley & Sons, Inc.,2014.
[19] J.J. van der Beek, O. Edfors, M. Sandell, S.K. Wilson, and P. O. Borgesson, “On
channel estimation in OFDM systems,” Proc. VTC 2015, pp. 815-819
[20] FarnooshTalaei and Xiaodai Dong, Senior Member, (2019),”Hybrid mmWave
MIMO-OFDM Channel Estimation Based on the Multi-Band Sparse Structure of
Channel”IEEE TRANSACTIONS ON COMMUNICATIONS, 4:247 – 267.
[21] Wei Ji, ChenhaoRen, Ling Qiu, Member, (2018),” Common Sparsity and Cluster
Structure based Channel Estimation for Downlink Massive MIMO-OFDM
Systems,”IEEE Transactions on Wireless Communications, 14(3):1356–1368.
[22] ShahabEhsanfar, Maximilian Matthe, MarwaChafii, (2018) ” Pilot- and CP-aided
Channel Estimation in MIMO Non-Orthogonal Multi-Carriers” IEEE Commun. Mag.,
52(2):97–105.
[23] Javad Mirzaei, Raviraj S (2019),”Semi-Blind Time-domain Channel Estimation for
Frequency-Selective Multiuser Massive MIMO Systems”, IEEE TRANSACTIONS
ON COMMUNICATIONS, IEEE Trans. on Signal Process., 54(3):884–893.
[24] Yuan Huang, Yigang He (2018),”Channel Estimation in MIMO – OFDM Systems
based on a new adaptive greedy algorithm”IEEE Transactions on Wireless
Communications, 15(4):2590-2603.
[25] H. Q. Ngo and E. G. Larsson, “EVD-based channel estimation in multicell multiuser
MIMO systems with very large antenna arrays,” in IEEE Int. Conf. Acoustics, Speech
and Signal Process. (ICASSP), March, 2012, pp. 3249–3252.
[26] K. Zarifi and A. B. Gershman, “Generalized correlation decompositionbased blind
channel estimation in DS-CDMA systems with unknown wide-sense stationary
noise,” IEEE Trans. on Sig. Proc., vol. 56, no. 11, pp. 5605–5617, Nov 2008.

[27] R. C. D. Lamare and R. Sampaio-Neto, “Blind adaptive MIMO receivers for space-
time block-coded DS-CDMA systems in multipath channels using the constant
modulus criterion,” IEEE Trans. on Commun., vol. 58, no. 1, pp. 21–27, January
2010.
[28] J. Lee, G.-T. Gil, and Y. H. Lee, “Channel estimation via orthogonal matching
pursuit for hybrid MIMO systems in millimeter wave communications,” IEEE
Transactions on Communications, vol. 64, no. 6, pp. 2370–2386, Jun. 2016.
[29] Y. Li, L.J. Cimini, and N.R. Sollenberger, “Robust Channel Estimation for OFDM
Systems with Rapid Dispersive Fading,” IEEE Transactionson Communications, vol.
46, pp. 902-915, July 2018.
[30] Charles U. Ndujiuba et.al: “Dynamic Differential Modulation of Sub-Carriers in
OFDM” Journal of Wireless Networking and Communications 2016, 6(1): 21-28.
[31] L. Dai, Z. Wang, and Z. Yang, “Spectrally efficient time-frequency training OFDM
for mobile large-scale MIMO systems,” IEEE J. Sel. Areas Commun., vol. 31, no. 2,
pp. 251–263, Feb. 2013.
[32] Y. Barbotin and M. Vetterli, “Estimation of sparse MIMO channels with common
support,” IEEE Trans. Commun., vol. 60, no. 12, pp. 3705–3716, Dec. 2012.

Research paper (channel_estimation)

More Related Content

What's hot (20)

Similar to Research paper (channel_estimation) (20)

Recently uploaded (20)

Research paper (channel_estimation)