Advanced hybrid algorithms for precise multipath channel estimation in next-generation wireless networks

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 14, No. 2, April 2024, pp. 1654~1664
ISSN: 2088-8708, DOI: 10.11591/ijece.v14i2.pp1654-1664  1654
Journal homepage: http://ijece.iaescore.com
Advanced hybrid algorithms for precise multipath channel
estimation in next-generation wireless networks
Kahina Rekkal1
, Sara Rekkal2
, Abdesselam Bassou3
1
Department of Electrical Engineering, Faculty of Technology, Salhi Ahmed University Center, Naama, Algeria
2
Department of Computer Science, University of Badji Mokhtar, Annaba, Algeria
Article Info ABSTRACT
Article history:
Received Aug 12, 2023
Revised Dec 5, 2023
Accepted Dec 18, 2023
Multipath channels continue to present challenges in wireless
communication for both 5G and 6G networks. A multipath channel is a
phenomenon in wireless communications where signals traverse from the
sender to the receiver along various paths. This end occurs due to the
reflection, diffraction, and refraction of signals of various objects and
structures in the environment. Such pathways can cause symbol interference
in the transmitted signal, leading to communication issues. To this end, our
paper proposes the integration of three algorithms: teaching-learning-based
optimization (TLBO), particle swarm optimization (PSO), and artificial
neural networks (ANN). This combination effectively analyzes and
stabilizes the transmission channel, minimizing symbol interference. We
have developed, simulated, and evaluated this hybrid approach for multipath
fading channels. We apply it to various coding schemes, including tail-biting
convolutional code, turbo codes, low-density parity-check, and polar code.
Additionally, we have explored various decoding methods such as Viterbi,
maximum logarithmic maximum a posteriori, minimum sum, and cyclic
redundancy check soft cancellation list. Our study encompasses new channel
equalization schemes and coding gains derived from simulations and
mathematical analysis. Our proposed method significantly enhances channel
equalization, reducing interference and improving error correction in
wireless communication systems.
Keywords:
Artificial neural network
Channel estimation
Particle swarm optimization
Polar code
Tail biting convolutional codes
Low density parity check
This is an open access article under the CC BY-SA license.
Corresponding Author:
Kahina Rekkal
Department of Electrical Engineering, Faculty of Technology, Salhi Ahmed University Center
BP- 66- Naama, Naama, Algeria
Email: rekkal@yahoo.fr, rekkal@cuniv-naama.dz
1. INTRODUCTION
As we delve into the ever-evolving landscape of wireless communication networks, pursuing higher
data rates, enhanced reliability, and the ability to support an increasing number of services has led us beyond the
realms of 5G and 6G networks [1]. This relentless march toward more advanced communication technologies
has unearthed numerous opportunities and challenges, and one of the core challenges lies in dealing with
multipath channels. Multipath channels are a ubiquitous feature of wireless communication, arising from the
complex interplay of signals as they bounce off, diffract, and refract through various environmental obstacles
and structures. The consequence of this phenomenon is the existence of multiple signal paths, each taking a
distinct route from the transmitter to the receiver. This multipath propagation engenders interference between
the symbols in the transmitted signal, leading to a phenomenon known as inter-symbol interference (ISI). ISI
poses a significant obstacle to achieving high data transmission rates, especially in scenarios where signals are
3
Department of Electrical Engineering, Faculty of Technology, Tahri Mohammed University Bechar, Algeria

Int J Elec & Comp Eng ISSN: 2088-8708 
Advanced hybrid algorithms for precise multipath channel estimation in … (Kahina Rekkal)
1655
subject to severe multipath propagation, such as in urban or indoor environments. Navigating the intricacies of
multipath channels and mitigating the challenges posed by ISI require innovative solutions that extend beyond
the capabilities of 5G and 6G networks.
Researchers and technology suppliers have been motivated to discover innovative methods to tackle
these challenges, as they substantially impact wireless communication quality and the radio spectrum's effective
utilization. The related works of researchers, including Sattiraju et al. [2], examined the performance of deep
learning (DL) architectures compared to the existing least-square (LS) channel estimation used in C-V2X. The
results of these studies show that the suggested DL architecture works better than the old C-V2X channel
estimation methods, especially when mobile speeds are high. Other researchers, including Mohammed et al. [3],
conducted studies evaluating the effectiveness of deep learning-based channel estimation (CE) in comparison to
conventional techniques such as least-square (LS) and minimal mean-square error (MMSE) estimators. Vidhya
and Kumar [4] proposed a hybrid methodology that integrates particle swarm optimization (PSO) and genetic
algorithms (GA) to address the challenge of channel estimation in multiple-input, multiple-output orthogonal
frequency division multiplexing (MIMO-OFDM) systems. Kadhim and Sallomi [5] proposed methodology
utilizes the sparsity of RIS channels to reduce the training complexity of the CNN model. Additionally, the
study employs PSO to optimize the hyperparameters of the CNN model. The simulation results indicate that the
suggested method achieves accurate channel estimation at a significantly reduced computational cost compared
to earlier methods.
Nahar et al. [6] introduced a new method for channel estimation (CE) in multi-carrier wireless
communication systems, specifically in orthogonal frequency division multiplexing (OFDM) systems and code
division multiple access (CDMA). The proposed approach combines the local search (LS) technique with the
particle swarm optimization algorithm. The simulation showed that the proposed CE of the MIMO-OFDM
system can significantly result in better BER performance compared with other techniques. Arora and Chawla
[7] combined two optimization techniques, the PSO and moth flame optimization (MFO) method, used to
optimize the performance rate in the same field. Sohail et al. [8] researched the precise and prompt collection of
extensive channel state information (CSI). They addressed an optimization challenge that can be tackled using
heuristic optimization methods such as genetic algorithms, particle swarm optimization, and differential
evolution.
Rahhal [9] utilized the PSO algorithm to identify the most optimal cluster members in a wireless
sensor network (WSN) based on a fitness function generated from the channel condition. This method
demonstrates the ability to rapidly and expeditiously achieve optimal cluster formation. Gondela et al. [10]
proposed a method for multiuser detection (MUD) in direct sequence-code division multiple access
(DS-CDMA) systems operating over generalized-K (GK) fading channels. The strategy utilizes the teaching-
learning-based optimization algorithm (TLBO) with a two-stage initialization (TSI) approach. This approach
demonstrates notable improvements in performance as the signal-to-noise ratio (SNR) and diversity order grow,
particularly in the presence of heavy-tailed impulsive noise. Ouali et al. [11] introduced a filter that consists of
two layers: a type2-fuzzy autoregressive filter and the training of membership function parameters. This filter
operates as a feedback mechanism. The second layer employs a teaching learning-based optimization algorithm
to adjust the settings of the type2-fuzzy adaptive filter. This adjustment aims to minimize the criterion function
and achieve signal reconstruction. The method surpasses denoising in terms of noise power levels and standards
views.
In this paper, we introduce an advanced hybrid algorithm that synergizes the capabilities of teaching-
learning-based optimization (TLBO), particle swarm optimization (PSO), and artificial neural networks (ANN)
for the precise estimation of multipath fading channel characteristics in wireless communication systems. This
novel methodology aims to optimize the parameter estimation of channel filters, significantly enhancing the
equalization process and effectively mitigating symbol interference in data transmission. Our method uses the
strong optimization abilities of TLBO and PSO to fine-tune the weights and biases of ANNs. This proposition
makes finding and fixing problems in wireless transmission signals easier. We rigorously evaluate the efficacy
of this hybrid algorithm using diverse datasets featuring signals encoded with various error-correcting codes,
such as turbo codes, low-density parity checks (LDPC), and polar encoders. These tests are conducted under
various channel conditions to ensure a comprehensive analysis.
The empirical results demonstrate that our method performs better than several contemporary
techniques, particularly precision and operational efficiency. This result is significant for new 5G and 6G
networks. Our study shows that hybrid optimization strategies can help a lot with the complex problems of
interference and symbol distortion in wireless signal transmission. This innovative approach opens new avenues
for enhancing communication reliability and quality in advanced telecommunication systems. The subsequent
sections of the paper are organized in the following manner: section 2 introduces sophisticated channel-coding
methods. Section 3 outlines the research methodology employed in this study. Section 4 shows the simulation
and results, comparing various situations. Finally, section 5 provides the conclusion.

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 14, No. 2, April 2024: 1654-1664
1656
2. ADVANCED CHANNEL CODING SCHEMES
Advanced channel coding is a pivotal strategy for ensuring reliable and efficient data transmission
across the air interface in 5G and 6G networks. This technique plays a crucial role in mitigating the impact of
noise, interference, and other forms of impairment that may compromise the integrity of transmitted data. By
integrating redundant information into the data stream, channel coding significantly reduces the likelihood of
data corruption. Implementing sophisticated channel coding methodologies is indispensable to augment the air
interface performance in 5G and 6G networks. Notably, high-tech error-correction codes like Polar codes and
low-density parity-check (LDPC) codes have shown to be very useful in modern communication systems.
These coding strategies facilitate the high-speed, low-latency connectivity essential for supporting emerging
applications and use cases within 5G and 6G network environments. Enhancing data transfer reliability and
efficiency through these methods is critical in advancing next-generation network technologies.
2.1. TBCC codes
In this study, we investigate using a specific error-correcting code within digital communication
systems: the tail-biting convolutional code (TBCC). TBCC is a subset of convolutional codes, which are
pivotal in adding redundancy to transmitted data by encoding the information into a sequence of symbols
[12]. Decoding these codes can be executed through several methods, with the Viterbi algorithm being a
prominent one. A trellis diagram is used in this algorithm to figure out the path metric value, which makes
decoding convolutional codes easier.
Adiono et al. [13] proposed a VLSI architecture for the reversed-trellis TBCC (RT-TBCC) method,
highlighting further advancements in this field. This innovative approach revises the traditional direct-
terminating maximum-likelihood decoding procedure to enhance the correction rate. The proposed
architecture is more computationally efficient and resource-effective than existing direct-terminating
maximum likelihood (ML) decoders. This architecture is evaluated through simulations, system-on-chip
implementations, and field programmable gate array (FPGA) synthesis. The focal point of this research is a
comparison of convolutional codes, specifically the TBCC, with a memory order of 𝑚 = 16, which adheres
to the LTE standards.
2.2. Turbo codes
In 1993, Claude Berrou and his colleagues introduced turbo codes, a revolutionary class of error
correction codes. These codes are notable for their capability to approach Shannon's theoretical channel
capacity limit with an unprecedented proximity of just 0.5 decibels (dB). This groundbreaking development
in information theory marked a significant step forward in achieving efficient data transmission with minimal
error rates. Turbo codes employ a dual structure of convolutional encoders, which is pivotal in forming these
codes. The decoding process of turbo codes is accomplished iteratively, utilizing the maximum a posteriori
probability (MAP) algorithm. These codes stand out for their recursive nature, allowing prior outputs and the
current input sequence to influence the output. This recursive output mechanism enhances the robustness and
efficiency of the codes, particularly in scenarios involving longer block lengths.
The exceptional performance of turbo codes has led to their widespread adoption in various
communication sectors. Their application ranges from deep space exploration missions, where reliable long-
distance communication is paramount, to terrestrial mobile communications, including 3G and 4G networks.
Turbo codes' high efficiency and low error rates make them an indispensable tool in these areas [14]. Recent
research has focused on developing and implementing turbo-coding in 5G technology. A notable project
involved designing and configuring a Turbo coder tailored explicitly for 5G networks [15]. This project
utilized advanced hardware description and verification tools such as Tanner, Verilog HDL (Xilinx), and
ModelSim for the individual configuration of the encoder and decoder. Implementing turbo coding is pivotal
in developing 5G technology, which demands high-speed, low-latency communication with minimal error
rates. The researchers' success in significantly enhancing data transmission reliability plays a crucial role in
the evolution of 5G networks, promising to revolutionize how we communicate and access information.
2.3. LDPC code
LDPC codes, introduced in the 1960s [16], represent a class of linear error-correcting codes
characterized by low encoding and decoding complexity. These codes are fundamentally based on sparse
matrices, allowing for efficient error correction in various communication channels, including wireless and
optical networks. The efficiency of LDPC codes has been instrumental in their integration into the 5G
network infrastructure. Notably, the 5G LDPC code, categorized as a quasi-cyclic LDPC (QC-LDPC) code,
is defined by two base graphs and is anticipated to play a significant role in the technological framework of
6G networks [17].

1657
Another variant, the rate-compatible raptor-like quasi-cyclic low-density parity-check (RL-QC-
LDPC) codes, offers remarkable performance and flexibility. Due to their enhanced error correction
capabilities, these codes have been widely accepted in 5G new radio (5G-NR) technical specifications [18].
The research by Zenkouar et al. [19] used the Group Shuffled Belief Propagation (GSBP) method to make
the GF(q) LDPC encoder and decoder work on an FPGA. These non-binary LDPC codes work exceptionally
well with short block lengths and have error correction rates close to the Shannon limit over GF (16). They
were modeled using VHSIC hardware description language and ModelSim 6.5.
The study of Ouakili et al. [20] introduced a new LDPC decoding algorithm combining normalized
min-sum and modified-weighted bit-flipping (NMSMWBF), exhibiting superior performance over traditional
NMS by 0.25 dB in bit error rate (BER) over AWGN channels. Razi et al. [21] also came up with two better
hard-decision algorithms: Adaptive gradient descent bit-flipping (AGDBF) and adaptive reliability ratio
weighted GDBF (ARRWGDBF). When implemented in real-time on digital signal processors, these work
better than soft-decision algorithms, achieving faster convergence and less processing time and memory
usage. For comparative purposes, this work examined an LDPC code based on 5G requirements [22]. The
decoding process employs the Min-Sum algorithm, highlighting its efficiency in terms of error correction
within the 5G network framework.
2.4. Polar codes
Polar codes, introduced by Arikan [23], represent a class of linear block codes that utilize a unique
coding scheme. This scheme involves transforming a set of independent and identically distributed (iid) bit
channels into channels with varying error rates, leveraging their polarizing properties. Due to their high-
performance capabilities, these codes are increasingly recognized as promising candidates for data
transmission in emerging 6G networks [24], [25].
The systematic turbo-polar code (STPC) with an early termination (ET) mechanism that Hamad
et al. [26] proposed is a notable development in the field of polar codes. This mechanism employs an ideal
scaling factor (SF) estimation technique, which has been shown to enhance the BER performance of STPC.
The SF estimation leads to a 1 dB improvement in BER over the traditional systematic polar code and a 0.3
dB improvement with a 64-bit code. The novel strategy involves using the estimated SF value at the second
component decoder and the decoded frozen bits as a criterion for stopping each decoding iteration. This
approach reduces the average number of iterations by half without increasing the system's complexity and
achieves BER results comparable to codes with a fixed number of iterations. The cyclic redundancy check
successive cancellation list (CRC-SCL) decoding method for Polar codes [27]–[29] was used for
comparisons of the CRC-SCL method, with a list size of 𝐿 =16 and a 16-bit CRC.
3. SYSTEM MODEL
Suggesting a hybrid method that integrates TLBO, PSO, and ANN to estimate the parameters of
channel filters. This approach aims to enhance equalization and minimize inter-symbol interference in
multipath fading channels. This technique seeks to exploit the advantages of each method: TLBO for
producing a robust initial estimate, PSO for optimizing the solution using a simulated social behavior model,
and ANN for refining the parameters with its adaptive learning capabilities. Our model consists of multiple
components, each playing a distinct role in addressing the challenges posed by multipath fading channels,
thereby improving the system's overall performance.
3.1. Multipath fading channel model
The multipath fading channel can be modeled as:
𝑦(𝑡) = ∑ ℎ . 𝑥(𝑡 − 𝜏 ) + 𝑛(𝑡) (1)
where 𝑦(𝑡) is the received signal, 𝑥(𝑡) is the transmitted signal, ℎ are the channel coefficients for the 𝑖
path, 𝑛(𝑡) is the noise in the channel, and 𝑁 is the number of paths.
3.2. ANN model for parameter estimation
The ANN is used to estimate the channel parameters like:
ℎ(𝑡) = ∑ 𝑎 𝛿(𝑡 − 𝜏 ) (2)
Based on the received signal:

 ISSN: 2088-8708
1658
𝑦(𝑡) = 𝑥(𝑡) ∗ ℎ(𝑡) + 𝑛(𝑡) (3)
The ANN might be structured as follows:
 Input layer: Takes in features derived from 𝑦(𝑡). These features could be various statistical properties or
transformations of the received signal, such as its power spectral density, autocorrelation, or any other
feature that captures the essence of the fading channel. Let's denote the input features as a vector 𝑥 ∈ 𝑅
and 𝑛 is the number of input features.
 Hidden layers: Multiple layers, each with a nonlinear activation function (like rectified linear unit (ReLU)
or sigmoid) denoted by 𝜎, is then applied:
𝑎( )
= 𝜎(𝑍( )
) (4)
Consider a network with 𝐿 hidden layers. Each layer 𝑙 applies a linear transformation followed by a
nonlinear activation function to its input. The linear transformation in layer 𝑙 can be represented as:
z( )
= W( )
a( )
+ b( )
(5)
where 𝑊( )
and 𝑏( )
are the weights and biases of layer 𝑙 and 𝑎( )
is the output from the previous layer
(or the input 𝑥 for 𝑙 = 1).
 Output layer: Provides estimates of the channel parameters.
3.3. PSO-TLBO hybrid optimization
The TLBO-PSO hybrid algorithm optimizes the weights and biases of the ANN. This process can be
abstracted as:
a. Initialize a population of candidate solutions (both teachers in TLBO and particles in PSO).
b. Evaluate the fitness of each candidate (how well the ANN with those weights or biases estimates the
channel parameters).
c. Apply TLBO and PSO operators to update candidates:
 TLBO: Based on the teaching phase and learning phase.
 PSO: Update the velocity and position of particles based on personal and global best positions.
d. Iterate until a stopping criterion is met (like a maximum number of iterations or a desired error threshold).
3.4. Equalization model
Once the channel parameters are estimated, they can be used in an equalization algorithm to mitigate
the effects of multipath fading. A new, widely linear equalizer has been proposed by Petitpied et al. [30]. A
common approach is using a linear or decision feedback equalizer. For example, a linear equalizer could be
modeled as (6):
𝑥(t) = ∑ 𝑤 . 𝑦(𝑡 − 𝑗) (6)
where 𝑥(t) is the estimated transmitted signal, 𝑤 are the equalizer coefficients, and 𝑀 is the number of
coefficients in the equalizer.
The hybrid PSO-TLBO-ANN pseudo-code has been applied in the manner shown as follows:
Input: Channel characteristics, SNR, Population size, Max Iterations, ANN architecture
Output: Optimized ANN weights, Performance metric
(*Initialization*)
Initialize the ANN with a given architecture and random weights
Initialize the population of particles (weights) and their velocities
Define PSO parameters: cognitive coefficient (c1), social coefficient (c2), and inertia
weight (w)
(*Optimization Loop*)
While (generation count < Max Iterations) do:
(*PSO Phase*)
For each particle do:
Calculate the particle's fitness using the ANN and channel model
Update the particle's personal best if the current fitness is better
Update the global best if the current fitness is better than the global best
End For
Update the particle's velocity based on its personal best, global best, and current

1659
velocity
Update the particle's position based on the new velocity
Apply the new weights to the ANN
Calculate the new fitness using the ANN and channel model
Update the personal and global bests if the new fitness is better
End For
(*Teacher Phase (TLBO)*)
For each particle (as a learner) do:
If (particle's performance is better than the mean performance of the population)
then:
Update the particle's position towards the best solution (teacher)
EndIf
End For
(*Learner Phase (TLBO)*)
Select another particle randomly
If (selected particle's performance is better) then:
Move the particle towards the selected particle
Else
Move the particle away from the selected particle
EndIf
End For
(*ANN Training Phase*)
Train the ANN using the best weights found by the hybrid PSO-TLBO
Evaluate the ANN's performance on a validation set
(*Update PSO parameters based on feedback*)
Adjust the PSO parameters (c1, c2, w) if necessary, based on the performance
(*Increment the generation count*)
generation count = generation count + 1
End While
(*Output the best solution*)
Return the best set of weights and the corresponding performance metric
The transmission chain in a multipath channel is described in several steps, as follows:
The transmission mechanism in the multipath channel depicted in Figure 1 can be decomposed into
numerous sequential steps:
 Data source: This is the starting point where the data to be transmitted is generated.
 Channel encoding: Techniques like LDPC or Polar coding encode the data for error correction
capabilities.
 Modulator: The encoded data is then modulated using a scheme such as quadrature phase shift keying
(QPSK) or quadrature amplitude modulation (QAM) to prepare it for transmission over the channel.
 Multipath channel: The modulated signal travels through a channel with multiple paths due to reflections
and scattering, which is typical in wireless communication environments.
 Channel estimation: Techniques PSO-TLBO-ANN are used to estimate the channel characteristics, which
is critical for the correct data reception.
 QPSK demodulator: The signal is then demodulated from the QPSK scheme back to a form that can be
decoded.
 Channel decoding: The demodulated data goes through channel decoding using techniques like Viterbi,
max-log-MAP, min-sum, SC, or CRC-SCL to correct any errors from the transmission process.
 Equalization: This stage makes up for the interference and signal distortion that the multipath channel
causes.
 Data sink: The data is received and processed at the destination.
The block diagram in Figure 2 involves a sophisticated hybrid algorithm combining particle swarm
optimization, teaching-learning-based optimization, and artificial neural networks to enhance the
performance of filters in addressing multipath fading channel issues in wireless transmission. The description
of this process is as follows:
a. Initialization
The process begins by generating an initial population of particles, each representing a potential
solution, and placing them randomly within the solution space. Initial settings of the algorithm are defined,
including the size of the population, the number of iterations, and parameters specific to the PSO, TLBO
methods, and the ANN configuration.
b. Performance assessment using TLBO
Each particle's performance is then evaluated. This end uses the TLBO method, which enhances the
particle's position in the solution space based on a teaching-learning paradigm. Concurrently, the ANN
method predicts potential damage locations in the wireless transmission signal.

 ISSN: 2088-8708
1660
c. Selection of the best particles
Based on their assessed performances, the best-performing particles are selected for participation in
the subsequent iteration.
d. Position update using PSO and performance reassessment
The chosen particles then undergo a position update using the PSO technique. This step optimizes
their placement in the solution space, aiming for a better approximation. After updating their positions, their
performances are reassessed using the TLBO and ANN methods. This reassessment focuses again on
predicting potential damage locations in the signal.
e. Iterative optimization
These steps are iteratively repeated for a predetermined number of iterations. The algorithm refines
the solutions during each iteration, gradually converging towards an optimal state.
f. Final outcome
Upon reaching the optimal state, the algorithm yields the most favorable positions for identifying
damage in the wireless transmission signal. It also provides insights into its overall performance, mainly
focusing on the speed and accuracy with which it can identify damages caused by the noisy channel.
Figure 1. Transmission chain in a multipath channel
Figure 2. Block diagram of PSO-TLBO-ANN proposed
The hybrid PSO-TLBO-ANN method synergistically integrates the key strengths of each approach
PSO's efficient optimization capability, TLBO's learning-based enhancement, and ANN's accurate prediction
ability to enhance the reliability and efficacy of signal transmission in a noisy, multipath fading channel.
Incorporating particle swarm optimization facilitates a flexible and responsive optimization procedure,
enhancing the system's ability to traverse intricate solution domains efficiently. Additionally, the integration
Data source
EX
Channel coding
LDPC, POLAR
MODULATOR
QPSK-QAM
Multipath channel
Channel estimation
PSO-TLBO-ANN
QPSK demodulator
Equalization
Channel decoding
Viterbi, MAX-Log -MAP,
min-sum,
SC, CRC-SCL
Data source
EX
Compute BER,
BLER
Generate an initial population of particles.
Evaluation of particles using the TLBO and ANN.
Selection of the best particles
(Intermediate generation).
Update the positions of particles by PSO. Evaluation of particles using the TLBO and ANN.
Positions of
particles >= Max
Result: the optimal position of damage in the wireless
transmission signal: 5G, 6G.
No
Yes

1661
of TLBO strengthens the learning component, guaranteeing continual improvement of the system over time
and utilizing ANN to aid in generating exact predictions, which is crucial for dependable signal processing in
demanding conditions. These combined qualities make the hybrid technique particularly suited for handling
the subtleties of multipath fading channels, where traditional methods might fall short.
4. RESULTS AND DISCUSSION
This study evaluates the Polar, Turbo, LDPC, and TBCC codes using multipath fading and additive
white Gaussian noise (AWGN) channels, together with the eight-phase shift keying (8PSK) modulation
scheme, as described in Section 2. This evaluation aims to determine each coding scheme's efficacy in
preserving signal integrity in difficult channel conditions. Table 1 presents a comprehensive list of the
parameters utilized, offering a clear overview of the experimental configuration and facilitating a thorough
comparison of the various coding schemes. The evaluation findings will provide valuable insights into
communication scenarios' most effective coding approach, particularly problems such as AWGN and
multipath fading.
Table 1. Simulation parameters
Parameters Specifications
Channel AWGN + Multipath
Modulation 8PSK,
Information block length (bits) 1024 bits
Code rate 2/3
Coding schemes TBCC Turbo LDPC POLAR
Decoding algorithm Viterbi Max-log-MAP
(20 iteration)
min-sum
(50 iterations)
CRC-SCL 16
Channel estimation PSO-TLBO-ANN
Channel equalizer ANN-TLBO-PSO
The results of our comprehensive simulation study are meticulously presented in Figures 3
through Figure 8, illustrating the bit error rate (BER) as a function of the signal to noise ratio (SNR). This
study demonstrates a significant enhancement in the performance of various coding methods, notably
polar, turbo, low-density parity-check (LDPC), and tail-biting convolutional codes (TBCC) when they are
integrated with a sophisticated hybrid algorithm combining particle swarm optimization, teaching-
learning-based optimization, and artificial neural networks. This hybrid approach, specifically designed to
estimate multipath channel effects, showcases an exceptional improvement in coding efficiency,
surpassing other traditional coding methods by approximately 0.5 decibels (dB), as detailed in Figure 4 to
Figure 8.
In particular, Figure 3 highlights the superior efficacy of polar codes when combined with cyclic
redundancy check successive cancellation list (CRC-SCL) decoders for a block length of K=1024 bits. This
setup outperforms almost all other coding schemes, achieving gains of 0.3 dB over turbo codes, 0.4 dB over
LDPC codes, and a remarkable 0.7 dB over TBCC. The graph also reveals that turbo codes have a slight edge
over TBCC, and they marginally surpass LDPC codes by about 0.1 dB.
Figure 4 through Figure 8 focus on the innovative hybrid PSO-TLBO-ANN algorithms employed to
predict the behavior of additive white gaussian noise (AWGN) channels affected by multipath fading. These
algorithms were tested using signals encoded with error-correcting codes, including TBCC, Turbo codes,
LDPC, and Polar encoders. According to Figure 4, when applied to polar codes, the hybrid algorithm
demonstrates superior performance over the turbo, LDPC, and TBCC codes, with gains of 0.3 dB, 0.4 dB,
and 0.8 dB, respectively. The simulations reveal a marginal yet noteworthy improvement in the performance
of turbo codes compared to LDPC codes, with a gain of 0.7 dB over TBCC. The enhanced performance can
be attributed to the TLBO and PSO strategies, which optimize the positioning of particles within the solution
space, coupled with using ANNs to accurately predict potential signal impairments, such as damages or
distortions, in wireless transmissions. Our findings suggest that employing robust encoding methods like
Polar, TLBC, and Turbo codes can be highly effective for 5G and 6G communication technologies,
particularly in mitigating issues related to noisy channels.
The hybrid PSO-TLBO-ANN algorithm, as evidenced by our simulations, demonstrates superior
capability in detecting and compensating for various transmission impairments, including intersymbol
interference, typically induced by noisy communication channels. This advancement marks a significant
step forward in enhancing the reliability and efficiency of next-generation wireless communication
systems.

 ISSN: 2088-8708
1662
Figure 3. Comparison of Polar codes, Turbo codes,
LDPC, and TBCC in multipath fading channels at a
rate of 2/3
Figure 4. Comparison of error-correcting codes
using hybrid PSO-TLBO-ANN estimation in
multipath fading environments
Figure 5. Comparison of Polar codes and Polar codes
with hybrid PSO-TLBO-ANN estimation in
Figure 6. Comparison of Turbo codes and Turbo
codes with hybrid PSO-TLBO-ANN estimation in
Figure 7. Comparison of LDPC codes and LDPC
Figure 8. Comparison of TBCC codes and TBCC

1663
5. CONCLUSION
This study presents the outcomes of a novel hybrid methodology that synergistically integrates the
strengths of teaching-learning-based optimization, particle swarm optimization, and artificial neural network
techniques. This innovative hybrid approach adeptly compensates for the limitations of each method by
leveraging their collective advantages. ANNs excel in discerning complex patterns and delivering precise
predictions. In contrast, TLBO and PSO are renowned for their efficiency in swiftly identifying optimal
solutions within a given solution space. This amalgamation enhances the accuracy in detecting transmission
signal impairments. The algorithm’s adaptability is further demonstrated by its successful application to
signals encoded with various error-correcting codes, including TBCC, turbo codes, LDPC codes, and Polar
encoders. This versatility suggests that the algorithm is not confined to a single signal type. In practical
terms, the hybrid PSO-TLBO-ANN approach significantly boosts the performance of Polar, Turbo, LDPC,
and TBCC codes in multipath channel estimations. It surpasses other coding techniques by approximately
0.5 dB. More specifically, the hybrid model yields a performance enhancement of 0.3 dB for Polar codes,
0.4 dB for LDPC codes, and 0.8 dB for TBCC compared to the Turbo codes. Conclusively, this research
paves the way for more reliable and precise communication in 5G and 6G wireless networks. The proposed
algorithm effectively mitigates the detrimental effects of noisy channels, marking a significant stride in
advancing wireless communication technologies.
REFERENCES
[1] P. Kamble and A. N. Shaikh, “6G wireless networks: vision, requirements, applications and challenges,” in 5th IEEE
International Conference on Advances in Science and Technology, ICAST 2022, Dec. 2022, pp. 577–581, doi:
10.1109/ICAST55766.2022.10039549.
[2] R. Sattiraju, A. Weinand, and H. D. Schotten, “Channel estimation in C-V2X using deep learning,” in International Symposium
on Advanced Networks and Telecommunication Systems, ANTS, Dec. 2019, vol. 2019-Decem, doi:
10.1109/ANTS47819.2019.9117972.
[3] A. S. M. Mohammed, A. I. A. Taman, A. M. Hassan, and A. Zekry, “Deep learning channel estimation for OFDM 5G systems
with different channel models,” Wireless Personal Communications, vol. 128, no. 4, pp. 2891–2912, Oct. 2023, doi:
10.1007/s11277-022-10077-6.
[4] K. Vidhya and K. R. Shankar Kumar, “Channel estimation of MIMO-OFDM system using PSO and GA,” Arabian Journal for
Science and Engineering, vol. 39, no. 5, pp. 4047–4056, Mar. 2014, doi: 10.1007/s13369-014-0988-8.
[5] J. Q. Kadhim and A. H. Sallomi, “Enabling deep learning and swarm optimization algorithm for channel estimation for low power
RIS assisted wireless communications,” International Journal of Interactive Mobile Technologies, vol. 17, no. 12, pp. 171–194,
Jun. 2023, doi: 10.3991/ijim.v17i12.39411.
[6] A. K. Nahar, M. M. Ezzaldean, S. A. Gitaffa, and H. K. Khleaf, “OFDM channel estimation based on novel local search particle
swarm optimization algorithm,” Review of Information Engineering and Applications, vol. 3, no. 2, pp. 11–21, 2016, doi:
10.18488/journal.79.2016.32.11.21.
[7] M. Arora and P. Chawla, “A hybrid optimization-based technique for channel estimation in OFDM system: a parametric
approach,” Wireless Personal Communications, vol. 128, no. 4, pp. 2571–2587, May 2023, doi: 10.1007/s11277-022-10060-1.
[8] M. F. Sohail, S. A. Ghauri, and S. Alam, “Channel estimation in massive MIMO systems using heuristic approach,” Wireless
Personal Communications, vol. 97, no. 4, pp. 6483–6498, Aug. 2017, doi: 10.1007/s11277-017-4849-0.
[9] J. S. Rahhal, “PSO for CWSN using adaptive channel estimation,” International Journal of Communications, Network and System
Sciences, vol. 06, no. 11, pp. 472–477, 2013, doi: 10.4236/ijcns.2013.611049.
[10] L. M. Gondela, V. K. Pamula, and A. K. Tipparti, “Multiuser detection over generalized-K fading channels using two-stage
initialized teaching learning based optimization,” in 2016 International Conference on Advances in Computing, Communications
and Informatics, ICACCI 2016, Sep. 2016, pp. 1309–1313, doi: 10.1109/ICACCI.2016.7732228.
[11] M. A. Ouali, M. Ghanai, and K. Chafaa, “TLBO optimization algorithm based-Type2 fuzzy adaptive filter for ECG signals
denoising,” Traitement du Signal, vol. 37, no. 4, pp. 541–553, Oct. 2020, doi: 10.18280/TS.370401.
[12] O. Işcan, D. Lentner, and W. Xu, “A comparison of channel coding schemes for 5G short message transmission,” 2016 IEEE
Globecom Workshops (GC Wkshps), Washington, DC, USA, 2016, pp. 1-6, doi: 10.1109/GLOCOMW.2016.7848804.
[13] T. Adiono, A. Z. Ramdani, and R. V. W. Putra, “Reversed-trellis tail-biting convolutional code (RT-TBCC) decoder architecture
design for LTE,” International Journal of Electrical and Computer Engineering (IJECE), vol. 8, no. 1, pp. 198–209, Feb. 2018,
doi: 10.11591/ijece.v8i1.pp198-209.
[14] Neha, “Implementation and performance analysis of convolution error correcting codes with code rate=1/2.,” in Proceedings -
2016 International Conference on Micro-Electronics and Telecommunication Engineering, ICMETE 2016, Sep. 2016,
pp. 482–486, doi: 10.1109/ICMETE.2016.111.
[15] S. Joseph, R. Kirubakkar, E. Kishore Kumar, and M. Mariammal, “Design and implementation of turbo coder for 5G technology,”
in 2021 7th International Conference on Advanced Computing and Communication Systems, ICACCS 2021, Mar. 2021,
pp. 934–938, doi: 10.1109/ICACCS51430.2021.9441690.
[16] P. Muangkammuen, W. Vongjampa, and P. Suthisopapan, “On the performance of rate-compatible LDPC codes under all-integer
quantization decoding,” Journal of Electrical and Computer Engineering, vol. 2022, pp. 1–8, Nov. 2022, doi:
10.1155/2022/3076517.
[17] T. Richardson and S. Kudekar, “Design of low-density parity check codes for 5G new radio,” IEEE Communications Magazine,
vol. 56, no. 3, pp. 28–34, Mar. 2018, doi: 10.1109/MCOM.2018.1700839.
[18] Y. Zhang, K. Peng, Z. Chen, and J. Song, “Construction of rate-compatible raptor-like quasi-cyclic LDPC code with edge
classification for IDMA based random access,” IEEE Access, vol. 7, pp. 30818–30830, 2019, doi: 10.1109/ACCESS.2019.2902884.
[19] F. Z. Zenkouar, M. El Alaou, and S. Najah, “GF(q) LDPC encoder and decoder FPGA implementation using group shuffled belief
propagation algorithm,” International Journal of Electrical and Computer Engineering (IJECE), vol. 12, no. 3, pp. 2184–2193,
Jun. 2022, doi: 10.11591/ijece.v12i3.pp2184-2193.
[20] H. El Ouakili, H. Touati, A. Kadi, Y. Mehdaoui, and R. El Alami, “New low-density-parity-check decoding approach based on

 ISSN: 2088-8708
1664
the hard and soft decisions algorithms,” International Journal of Electrical and Computer Engineering (IJECE), vol. 13, no. 2,
pp. 1639–1646, Apr. 2023, doi: 10.11591/ijece.v13i2.pp1639-1646.
[21] M. Razi, M. Benhayoun, A. Mansouri, and A. Ahaitouf, “An improvement and a fast DSP implementation of the bit flipping
algorithms for low density parity check decoder,” International Journal of Electrical and Computer Engineering (IJECE),
vol. 11, no. 6, pp. 4774–4784, Dec. 2021, doi: 10.11591/ijece.v11i6.pp4774-4784.
[22] M. Syarhan Idris, D. Mohd Ali, N. Idora Abdul Razak, A. Idris, and H. Ahmad, “Performance analysis of NOMA using different
coding techniques,” Journal of Physics: Conference Series, vol. 1502, no. 1, Mar. 2020, doi: 10.1088/1742-6596/1502/1/012002.
[23] E. Arikan, “Channel polarization: A method for constructing capacity-achieving codes for symmetric binary-input memoryless
channels,” IEEE Transactions on Information Theory, vol. 55, no. 7, pp. 3051–3073, Jul. 2009, doi: 10.1109/TIT.2009.2021379.
[24] L. Mucchi et al., “Signal processing techniques for 6G,” Journal of Signal Processing Systems, vol. 95, no. 4, pp. 435–457, Feb.
2023, doi: 10.1007/s11265-022-01827-7.
[25] C. Kestel, M. Geiselhart, L. Johannsen, S. ten Brink, and N. Wehn, “Automorphism ensemble polar code decoders for 6G
URLLC,” Arxiv.org/abs/2303.01235v1, Mar. 2023.
[26] A. A. Hamad, M. T. Gatte, and L. A. Abdul-Rahaim, “Efficient systematic turbo polar decoding based on optimized scaling factor
and early termination mechanism,” International Journal of Electrical and Computer Engineering (IJECE), vol. 13, no. 1,
pp. 629–637, Feb. 2023, doi: 10.11591/ijece.v13i1.pp629-637.
[27] H. Suhaimi, R. Mohamad, D. M. Ali, and E. Abdullah, “Cyclic redundancy check-aided successive cancellation-based polar
decoders,” Indonesian Journal of Electrical Engineering and Computer Science (IJEECS), vol. 32, no. 2, pp. 811–818, Nov.
2023, doi: 10.11591/ijeecs.v32.i2.pp811-818.
[28] L. Zhang, A. Cao, J. Qiao, and Y. He, “A CRC-aided SR-SCL decoding algorithm for Polar codes,” 2019 IEEE MTT-S
International Wireless Symposium (IWS), Guangzhou, China, 2019, pp. 1-3, doi: 10.1109/IEEE-IWS.2019.8804046.
[29] C. Miao, Z. Shuang, and Z. Shengmei, “Multiple CRC-aided variable successive cancellation list decoder of Polar codes,”
Journal of China Universities of Posts and Telecommunications, vol. 24, no. 2, pp. 83–88, Apr. 2017, doi: 10.1016/S1005-
8885(17)60202-4.
[30] T. Petitpied, R. Tajan, G. Ferre, P. Chevalier, and S. Traverso, “A new widely linear equalizer based on expectation propagation
for faster-than-nyquist,” 2019 IEEE Global Communications Conference (GLOBECOM), Waikoloa, HI, USA, 2019, pp. 1-6, doi:
10.1109/GLOBECOM38437.2019.9014275.
BIOGRAPHIES OF AUTHORS
Kahina Rekkal Born in Bechar, Algeria, the author earned her Dipl.Inf.-Ing.
degree from the University of Bechar in 2009, a master’s degree in 2012, and a Doctorate in
2018, all from the same university. Currently serving as an assistant professor at Salhi Ahmed
Naama University Center in Algeria, her professional focus encompasses channel coding,
digital signal processing, optimization of trellis-coded modulation schemes, genetic
algorithms, communication over multipath and fading channels, orthogonal frequency-division
multiplexing (OFDM), channel equalizers, renewable energies, and Arabic natural language
processing. She can be contacted at email: rekkal@cuniv-naama.dz.
Sara Rekkal Born in Bechar, Algeria, the author obtained her master’s degree in
computer science from the University of Bechar in 2015 and completed her Doctorate in 2019 at
the University of Ahmed Ben Bella Oran 1. She is an assistant professor at Badji Mokhtar
Annaba University in Algeria. Her primary areas of expertise include web services, multi-criteria
decision support, embedded systems engineering and parallelism, metaheuristic algorithms, and
natural language processing. She can be contacted at email: rekkalsarah08@gmail.com.
Abdesselam Bassou Abdesselam Bassou was born in Bechar, Algeria. He
received the Dipl.El.-Ing. Degree from the University of Tlemcen, Algeria, in 1997, and a
master’s from the University of Sidi Bel Abbes, Algeria, in 2000. He obtained his doctorate
from the University of Sidi Bel Abbes, Algeria in 2006. He is a professor at the University of
Bechar, Algeria. His research interests include digital signal processing, Turbo encoding
schemes, and iterative decoding over fading channels. He can be contacted at email:
bassou.abdesselam@univ-bechar.dz.

Advanced hybrid algorithms for precise multipath channel estimation in next-generation wireless networks

More Related Content

Similar to Advanced hybrid algorithms for precise multipath channel estimation in next-generation wireless networks (20)

More from IJECEIAES (20)

Recently uploaded (20)

Advanced hybrid algorithms for precise multipath channel estimation in next-generation wireless networks