Machine Learning Based 5G Network Channel Quality Prediction

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 12 | Dec 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 630
Machine Learning Based 5G Network Channel Quality Prediction
Riya Sharma1, Dr. Pramod Sharma2
1Student, Dept. of ECE, RCERT, Jaipur, Rajasthan, India
2Professor, Dept. of ECE, RCERT, Jaipur, Rajasthan, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract –
Channel quality feedback is very important for operation of 4G or 5G wireless complex because it allocate user equipment (UE)
connections, transmission scheduling, or control over modulation or rate of data spread over wireless connection. Though, if
comments like this occur frequently and the number of UEs withinacell islarge, channel maybeoverloadedbysignalingmessages,
reducing throughput or loss of data. Therefore, optimizing this signaling process is an important challenge. This thesis focuseson
channel quality indicator (CQI) report irregularly transmitted from the UE to base station, and providesamechanism tooptimize
reporting procedure with aim of reducing signaling overhead or avoiding overload of connected channel and detect channel
quality and reconstruct node, For this purpose, machine learning techniques are applied to predict the stability of the channel.
Implemented CNN and SVM Algorithm for channelqualityestimation. AndproposedtheutilizationofParticleSwarmOptimization
(PSO) in Convolutional Neural Networks (CNNs), which is one of the basic methods in deep learning. TheuseofPSOonthetraining
process aims to optimize the results of the solution vectors on CNN in order to improve the recognition accuracy. This Simulation
Has Performed on the MATLAB Simulation. The simulation results show that all provide highpredictionaccuracy when compared
to traditional methodologies.
Keywords - Channel Quality Prediction, 5G Network, Machine Learning, CNN,SVM
I. INTRODUCTION
Machine learning is made up of algorithms that can learn from data and make predictions based on what they have learned.
These kinds of algorithms make predictions or decisions by building models based on the information they are given, rather
than by following a set of rules. This set of algorithms has been used well in many different fields, such as computer security,
bioinformatics, computer vision, medical diagnostics, and search engines, to name a few. All of thesesystemshaveonethingin
common: they can automatically look at databasedata tofindactionableinsightsandmakedecisionsbasedonthatdata.Mobile
networks are known for being hard to understand, and it seems likely that the new 5G communication systems will be even
harder to understand. They need to be able to handle a growing number of situations that can't be fully communicated with
mobile systems of today. Some examples of these scenarios are multiple deployments of powerful power lines, intelligent
transportation systems, low-latency connections, and networks' company[1]. To deal with this level ofcomplexity, weneedto
come up with sophisticated ways to look at 5G data. In order to make decisions, these methods need to be able to gather
information, reduce the number of people needed to run these communicationnetworks,cutdownontheamountofwork that
comes with managing networks, and predict how users and networks will act in the future.
Channel Quality Indicator-
The Channel Quality Index, which is also written as CQI, is a measure of how well information can be sent over a wireless
channel. A CQI can be a value or many values, and it stands for a metric that measures how good a certain channel is. Most of
the time, a CQI with a high charge means that the channel also has a high charge, and vice versa. Use performance indicators
like the signal-to-noise ratio (SNR), the signal-to-interference plus noise ratio (SINR), the signal-to-noise plus distortion ratio
(SNDR), etc. to figure out the Channel Quality Index (CQI). By figuring out these or other values for a channel, you can then use
those values to figure out the CQI for that channel[2]. The CQI of the channel can be affected by the type of transmission
(modulation) used by the communication system. For example, a communication company that uses multi-input distribution
code (CDMA) can use a wider range of CQIs than one that uses orthogonal division multiplexing (OFDM). In more complicated
communication systems, like those that use multi-channel input (MIMO) or space coding, the CQI may also dependonthetype
of receiver being used. Things that can be taken into account in CQI include the failure to carry out thedemonstration,Doppler
shift, the evaluation of the information channel, interference, and so on.

II. RELATED WORK
Sihem Bakri et.al. 2020 [1] Channel quality feedback is crucial for the operation of 4G and 5G radio networks, as it allows
controlling User Equipment (UE) connectivity, transmission scheduling, and the modulation and rate of the data transmitted
over the wireless link. However, when such feedback is frequent and the number of UEs in a cell is large, the channel may be
overloaded by signaling messages, resulting in lower throughput and data loss. Optimizing this signaling process thus
represents a key challenge. In this paper, we focus on Channel Quality Indicator (CQI) reports that are periodically sentfroma
UE to the base station, and propose mechanisms to optimize the reporting processwiththeaimofreducingsignalingoverhead
and avoiding the associated channel overloads, particularly when channel conditionsarestable.Tothisend,weapplymachine
learning mechanisms to predict channel stability, which can be used to decide iftheCQIofa UEisnecessarytobereported,and
in turn to control the reporting frequency. We study two machine learning models for this purpose, namely Support Vector
Machines (SVM) and Neural Networks (NN). Simulation results show that both provide a high prediction accuracy, with NN
consistently outperforming SVM in our settings, especially as CQI reportingfrequencyreducesobtainedthebestASE usingthis
method.
Lubov Berkman et.al (2019)[2] the change in channel capability appears as a change in a number of objectives, which are
proposed to determine the value of state-of-the-art control technology. It is recommended to use the gradient prediction
method to predict the state of the channel. Analyze the parameters that characterize the separate channel state. The control
element needs to consider the characteristics of the control channel and put in place to improve the efficiency of the control.
Consider how to evaluate the quality of a communication channel. We define the algorithm that is proposed to be used on a
network with packet packets during the processing of access to limit the network bandwidth. According to the general and
partial details of the improvement, consider the possibility of choosing the best method. Analyze the measurement of quality
and service standards. It is very convenient to use a continuous measurement method and a clear output method to measure
the load. A network exchanges packages when measuring service quality indicators (number of reports, expectedtimetostart
service and τ). π) It is very convenient to use the method of directly counting the number of reports.
V. A. Babkin et.al (2019)[3] In order to ensure the quality of traffic flowina communicationnetwork,itisnecessarytoensure
the value of the quality index within an acceptable time frame. One of these indicators is the traffic transmission rate reported
in the traffic data processing. By checking whether the user's traffic managementfile matchestheconfigurationfile specifiedin
the configuration file, the quality control value can be kept within a single value. of the traffic, thus maintaining the quality of
the user’s traffic.
Hesham M. Elmaghraby et al. (2018)[4] this paper solves the problem of channel distribution for femtocells that share the
common use of macrocells. The program problem of the femto base camp (FBS) is presented in the form of a Restless Armed
Rogue (RMAB) system. Our goal is to select a branch / channel that optimize the amount of expected reduction rewardoveran
indefinite period of time, while minimizing the interference caused by cell division channel distribution. Instead of directly
monitoring the actual channel quality, we use a cellular user feedback called the Channel Quality Index (CQI). In general, the
RMAB problem is a PSPACE problem. In order to estimate the available channels in the FBS, we propose an indexing strategy
with low inference difficulty, called the Wit Le average index. Finding a closed channel reservation solution often means that
there are closed channel reservations that have an active program but are based on partial channel informationintheCQI.We
also highlight the benefits of a referral policy over a short -sighted policy.
III .PROPOSED SYSTEM
5G networks can give users a better experience because they have more capacity or better management, but they also need
more accurate channel predictions than older mobile networks did. In this thesis, a machine learning method, including the
CNN and SVM algorithms, was suggested as a way to predict the Channel Quality Index (CQI). Reflection,diffraction, andsignal
scattering are the three things that, in a typical cellular communication scenario, When compared to LTE, the physical layer
resources available in a 5G network are more plentiful but also more difficult. As a result, algorithms for scheduling that are
more flexible and dependable, in addition to CQI values that are more accurate, are of utmostsignificancefortheadvancement
of NR. networks. CQI creation and reporting have traditionally been carried out by favor the idea of delaying the timetable,
which will inevitably lead to a decrease in the system's overall performance. Utilizing tools of prediction such as The CQI's
accuracy can be improved with the help of deep learning techniques. which is finally something that helps out the NR system.
0+/8A. The Prediction of CQI Through the Use of Deep Learning Algorithms At the moment, we are concentrating on the
wideband CQI, which is a positive integer. between 1 and 15. Considering that the CQI is merely a discrete value while UEs are
continuous, the It is not possible to merely train because conditions are constantly shifting. a component that can predict the
actions of any user based solely on their history importance of the CQI. In the event that a user is going in the direction of the

base station at whereas at other times it is moving away from the BS, it is moving away from the BS this time. Because of this,
the module won't know how to respond to the user's actions. Afterward, all it does to learn is how to mimic the conduct it has
already seen. This being said, in this manner, the training module will only require a few steps to complete, and the result
indicates that there was an apparent delay in the CQI result when compared to the actual value that has been reported. In
addition to this, the actions taken by users ar vary, hence the BS ought to have different models for each single variation. users
who connected themselves to it.
Figure1 . Illustration of CQI prediction module.
At the physical layer of a 5G network, there are not only more and better resources than in an LTE network, but there are also
more of them. Because of this, it is very important to improve NR networks by making their scheduling algorithms more
flexible and reliable and by giving them more accurate CQI values.OverthecourseofCQI'shistory,itsproductionandreporting
have been known to cause schedule delays, which in turn have led to a drop in system performance. We can improve the
accuracy of the CQI by using prediction methods like deep learning algorithms, whichisgoodfortheNRsysteminthelongrun.
A downlink scheduler, which is also called a MAC Scheduler, is part of the NR system's medium access (MAC) layer. This
scheduler's job is to get the user's personal schedule information when the user connects to the base station. Thisinformation
includes the CQI as well as Quality of Service messages and buffer status reports (BSRs) sent by the Radio Link Control (RLC)
layer (QoS). The scheduler will then pick a user to represent each RB based ontheinformationthathasalreadybeen given. The
value of the CQI is used to figure out the MCS, which then gives informationaboutthe TransportBlock Size. When RBsaregiven
out to users, this is what happens (TBS). There are limits on how much data can be sent during this time, which are set byboth
the MCS and the TBS. Users have to give the CQI at every time interval that has already been set. The base station had to send
the CQI before it could get the feedback, so there will be a delay between when it asks for the CQI feedback and when it gets
it[5-7].
At the physical layer of a 5G network, there are not only more and better resources than in an LTE network, but there are also
more of them. Because of this, it is very important to improve NR networks by making their scheduling algorithms more
flexible and reliable and by giving them more accurate CQI values.OverthecourseofCQI'shistory,itsproductionandreporting
have been known to cause schedule delays, which in turn led to a drop in system performance. Wecanimprovetheaccuracyof
the CQI by using prediction methods like deep learning algorithms. This is good for the NR system in the long run.
A downlink scheduler, also called a MAC Scheduler, is part of the NR system's medium access (MAC) layer. When a user
connects to the base station, it is the job of this scheduler to get the user's personal schedule information.
This information includes the CQI as well as Quality of Service messages and buffer status reports (BSRs) that are sent by the
Radio Link Control (RLC) layer (QoS). The scheduler will then choose a user to represent each RB, taking into account the
information that was given before. The value of the CQI is used to calculate the MCS, which, inturn,tellsusabouttheTransport
Block Size.
When RBs are given to users, this is what happens (TBS)[8]. The MCS and the TBS both put limits on how much data can be
sent at this time. Users are required to give the CQI at every predetermined time interval. The base station had to send the CQI
before it could get the feedback, so there will be a delay between when it asks for the CQI feedback and when it finally gets it.

Figure. 2 Flow Diagram
IV SYSTEM DESCRIPTION
Using machine learning to make a system that can predict the future To make a prediction system, a machine learning
algorithm goes through the next two steps.
a) Training phase: Seventy percent of the feature vectors and their labels are used to train the classifier (which reflect the
actual classes). During the training step of machine learning, a function thatmapsinputs(called"featurevectors")tooutputsis
made (labels). Then, this function is used to put new vectors into groups. At this point, the NN algorithmcanhandle bothlinear
and nonlinear functions, while the SVM method only learns linear functions.
b) Test and validation phase: In this step, we use the remaining feature vectors (30 percent ). It involves comparing the
predicted classes for these vectors with the labels that have already been given to them.
Particle Swarm Optimization (PSO)- After that, this method can be used to improve the way CQI data messages are sent,
which will lead to less signalling overhead in the long run. And showedhowParticleSwarmOptimization(PSO), whichisoneof
the most important deep learning techniques,canbeused inConvolutional Neural Networks(CNNs).Duringthetrainingphase,
PSO is used to improve the accuracy of the recognition process by making the results of the CNN-generatedsolution vectorsas
good as possible.
Imagine a network with N nodes of user equipment and one base station. All of the nodes can talk to each other (BS).Aspart of
these conversations, the Channel Quality Index (CQI) of the communication frequency bands is given so that the conditions of
those frequency bands can be looked at. This makes it easier for people totalk toeachotherina waythatworksbetter.Because
of this, the signal-to-noise ratio (SNR) for each subcarrier is a good choice of CQI, and it will be used as such for the rest of this
investigation. In reality, the CQI is either a 4-bit value (for 5G) or a 5-bit number, and it encodesthechannel gainsaswell asthe
modulation and coding scheme being used (MCS). Sub bands are smaller, morespecificpartsofeachfrequencybandwidththat
are made as it is divided further. After that, each sub-band is divided into physical resource blocks(PRB),whicharemadeupof
sub-carriers in the end (SC)
Table 1- Network Parameters
Parameters Value
Number of nodes 150
Area length(m) 400
Area width (M) 400
Expected sensors readings 30
Size of data buckets (bytes) 400

5G Dataset with Channel and Context Metrics-An example of a 5G trace file froma largemobilenetwork providerinIreland,
which is part of the United Kingdom. Ireland is home to the headquarters of the company that runs this service. The data set
was made by putting together two differentapplicationpatternsandtwodifferent mobilitypatterns—thestaticpatternandthe
vehicle pattern (video streaming and file download). During the data collection process, the following are the most important
performance indicators of the client-side cellular network: (KPIs). Some of the key performance indicators (KPIs) that are
included here are throughput, indicators about cells, metricsaboutcontext,andmetricsaboutchannels.Awell-knownAndroid
programme called G-NetTrack Pro was used to take these measurements. It was made for network monitoring and doesn't
require the device to be rooted in order to work. As far as we know, this is the first time that informationaboutthethroughput,
channel, and context of 5G networks has been made public[9].
Figure 3. dataset
Both a real-time 5G production network dataset and a MATLAB modelling framework for large-scale multi-cell 5G networks.
Now that the 5G/mm wave module for the ns-3 mm wave network simulator is available, we will be able to learn more about
how adaptive clients in 5G multi-cell wireless situations come to their conclusions.
The main goal of our framework is to give end users more information, such as a variety of metrics that are only relevant to
users who are connected to the same cell10-12]. This will give end users access to information they couldn'tgetanyotherway
about the environment of the base station (eNodeB or eNB) and the scheduling principle. We make it possible for other
academics to look into this interaction by letting them use our technology to make their own fake datasets.
V SIMULATION RESULTS
When judging the quality of the communication model as a whole, it is important to take into account the signal-to-noiseratio.
The Modulation Schemes and the Common Quality Index. It is very important to understand how each of these parts fits into
the whole. The correlation quality index (CQI) and the signal-to-noise ratio (SNR) are related, according to the results of this
research project (CQI). It was decided that this deal would give a better data rate than other partnerships that were alreadyin
place. When judging the overall quality of a communication model,itisimportanttolook attheSignal-to-NoiseRatio(SNR),the
Channel Quality Index (CQI), and the Modulation Schemes. It is very important to understand how each of these parts fits into
the whole. The correlation quality index (CQI) and the signal-to-noise ratio (SNR) are related, according to the results of this
research project (CQI). It was decided that this deal would give a better data rate than other partnerships that were alreadyin
place.

Figure.4 parameters initialization
Fig 5 window showing the parameters initialization, the initial parametersgivennumberof nodes,area,widths,packetsrates
and expected sensor reading.
Figure 5 node deployed in the network

Fig 6 showing the node deployed in the network, in this simulation 150 nodes deployed in the network
Figure.6 network creation and node deployment of nodes
The figure 7 depicts the initialization of parameters in this network, which has 150 nodes and a length and width of 400m.and
size of the data bucket 400.
Figure.7 check quality of node in the network
After the network has been set up, all of the nodes are spread out randomly across the network. From there, packets are sent
from one node to the next. All over the blue node, you can see where the data nodes are. This node's job is to make it possible
for information to move from one node to another. Figure 8 shows us this.

Figure. 8 check quality of node in the network
Some nodes may be weak and unable to send packets from one node to another while others are sending packets from one
node to another. Machine learning was used to check the quality of the node, and theweak nodeinthenetwork isshownbythe
red node. Figure 9 shows the channel quality indicator for this network, which was based on the channel quality dataset.
Figure.9 weak nodes

Fig 10 showing the weak nodes in the network those node not full fill therequirementofthechannel qualityconsiderasa weak
node in the network.
Figure.10 reconstruct node in the network
After applying a machine learning algorithm, the green node represents the reconstructed node in the network,showinginfig
11
Fig.11-performance of network
Fig. 12 showing the performance of the network in terms of PSNR, BER, MSE, accuracy .
CQI Value Ranges=On a scale from 0 to 30, where 30 means the channel is the best and 0 means it is the worst, 30 is the best
rating that can be given to the channel. The size of the transport blocks that the network uses to send data varies and is based
on what is reported to the EU. When user equipment (UE) transmits a high CQI to the network, the network responds by
sending larger block sizes. When the opposite is true, smaller block sizes are used to send information.

Even though the user says that the CQI is low, it is still possible that the network is sending a lot of data. If the UE has a CRC
error, it is likely that it won't be able to figure out what the information is. Because of this, the network will have to send it
again, which is a waste of the radio resources that are available.
What should a user do if the actual channel quality is low even though the UE claims to have a high CQI? In this case, if the
network sends a big transport block size, it's more likely that the UE won't be able to decode it, which would cause a CRC error
on the UE side of the communication. This would mean that the network would have to send it again, whichwouldwasteradio
resources. In this case, the CQI value will tell the network that each transport block needs to carry a lot of data.
Fig. 12- Performance Graph
Training Results
 numbatches = 0.0200
 epoch 1/1
 Elapsed time is 0.157340 seconds.
 Training error = 100%
 TIME2 = 0.426
 Testing error = 150 Nodes
 TIME1 = 5.4759
Performance Evaluation Model-Bit Error Rate (BER): The Bit Error Rate (BER) is a way to figure out how many bit errors
happen in a certain amount of time. Divide the total number of bit errors that happened during this time period by the total
number of bits that were sent. This is called a bit error rate (BER) (BER).BERisalmostalwaysgivenasa percentage lower than
units when measuring performance. To figure out how likely it is thata bitwill gowrong,youmustfirstfigureoutthepredicted
bit error rate. The bit error rate is a measurement that can stand in for the chance that a bit error will happen. This estimate is
correct for times that are longer than one second and for errors that involve more than one bit. The formula can be used to
show that the BER is a function of Eb/N0 for both QPSK and AWGN modulation.
BER = erfc ( / )
There are several ways to figure out how good a picture compression is. Two of these ways are the mean square error (MSE)
and the peak signal-to-noise ratio (PSNR). The MSE statistic shows the total squared error betweenthecompressed imageand
the original image, while the PSNR statistic shows the error in its worst form. As the total number of errors goes up, the MSE
value will start to go down.

Peak signal-to-noise ratio (PSNR) - It is the signal-to-noise ratio that shows how well a signal can be shown at its highest
possible value (power) compared to how much noise affects its accuracy (PSNR). This percentage is shown as a number.
PSNR = 10log10( )
10.log10 ( )
20.log10(MAXI)-10.log10(MSE)
The MAXI command shows the maximum value that can be used for a given pixel in an image. Using 8 bits per sample to show
the pixels, this value is the same as the number 255. When samples are encoded with linear PCM andB bitspersample,MAXI is
often the same as 2B1.
Mean squared error (MSE) - MSE or MSD is a way to measure the average square of an estimator's mistakes or the average
square of the difference between what was estimated and what was measured. This number is also called the "mean squared
error," the "mean squared deviation," and other names.
MSE =
MSE} = mean squared error
{n} = number of data points
Y{i} = observed values
{Y}{i} = predicted values
Table 2- Comparison result with exiting work
Technique Optimization Technique Accuracy (%)
Proposed System
ConvolutionNeural Network Particle Swarm Optimization (PSO) 99.41
Support Vector Machine 95.21
Existing System Support Vector Machine 92.86

Table 3- Performance of proposed algorithm
Technique BER Mean Square
Error
Peak Signal-To-Noise
Ratio
True Positive
Rate
Proposed
System
Convolution
Neural Network
0.41 1.05 47.8 0.12
V CONCLUSION
The routine sending of CQI reports, which give information about the quality of the channels in 4G and 5G mobile networks,
adds to the amount of signalling that needs to be done to keep the networks running. The CQI reports tell us how good the
channels in 4G and 5G mobile networks are. We looked into a lot of different ways to lower the total cost. We think that if we
think about how stable the channel is, we won't have to send CQI signals whenit'snot necessarytodoso.Toputitanother way,
we will send out fewer CQI reports when the value of the CQI doesn't change much over time. This means that the channel is
working as it should. To do this, we put a lot of time and effort into making machine learning-based strategies that only need
CQI data as an input. The goal of this exercise was to figure out how to make accuratepredictionsabouthowthechannel would
behave. So, the way our mechanisms work is in line with the standards and doesn't need any cross-layer or other outside
information, like where the users are or how they move around the environment.
In this particular situation, we looked at how well Support Vector Machines (SVM) and Convolution Neural Networks (CNN)
could predict the results. We also looked into whether there is a link between how accurate our predictions are and how
quickly we get new information. The results of our tests showed that neural networks always did better than SVMs in every
situation we tested them in. During this investigation, we spent most of our time trying to figure out how well the ML systems
we were looking at could predict the future. The next step is to start a more in-depth study of how the proposed method and
processes affect the 5G network slice management architecture that we proposed in our earlier study. The next step in the
process will be to do this. At the moment, our main goal is to figure out how to improve signaling and what effectour ideas will
have on how available resources are used and how well 5G networks
References
1. Sihem Bakri∗, Maha Bouaziz∗, Pantelis A. Frangoudis‡, and Adlen Ksentini∗ Channel stability prediction to optimize
signaling overhead in 5G networks using machine learning Sihem Bakri∗,Maha Bouaziz∗,PantelisA.Frangoudis‡,and
Adlen Ksentini∗ 978-1-7281-5089-5/20/$31.00 ©2020 IEEE

2. Lubov Berkman;Olha Tkachenko;Larysa Kriuchkova;Oksana Varfolomeeva DeterminationofCriteria forChoosingthe
Best Ways and Indicators of Service Quality in Infocommunication Networks 2019 3rd International Conference on
Advanced Information and Communications Technologies (AICT) Year: 2019
3. V. A. Babkin;E. P. Stroganova Integral Quality Control of Service Profile of Traffic 2019 Systems of Signals Generating
and Processing in the Field of on Board Communications Year: 2019
4. Haifeng Guo Model Predictive Control on Dual-Channel Closed-Loop Supply Chain Model andItsBullwhipEffect2018
Eighth International Conference on Instrumentation&Measurement,Computer,CommunicationandControl (IMCCC)
Year: 2018
5. Pengfei Li;Yun-Bo Zhao;Yu Kang Integrated Channel-Aware Scheduling and Packet-Based Predictive Control for
Wireless Cloud Control Systems IEEE Transactions on Cybernetics Year: 2020
6. Chi-Chuan Ho;Bo-Hong Huang;Meng-Ting Wu;Tin-Yu Wu Optimized Base Station Allocation for Platooning Vehicles
Underway by Using Deep Learning Algorithm Based on 5G-V2X 2019 IEEE 8th Global Conference on Consumer
Electronics (GCCE) Year: 2019
7. Sahrish Khan Tayyaba;Hasan Ali Khattak;Ahmad Almogren;Munam Ali Shah;Ikram UdDin;IbrahimAlkhalifa;Mohsen
Guizani 5G Vehicular Network Resource Management for Improving Radio Access Through Machine Learning IEEE
Access Year: 2020
8. Joel Shodamola;Usama Masood;Marvin Manalastas;Ali Imran A Machine Learning based Framework for KPI
Maximization in Emerging Networks using Mobility Parameters 2020 IEEE International Black Sea Conference on
Communications and Networking (BlackSeaCom) Year: 2020
9. V.A. Babkin;E.P. Stroganova Principles of Indicators Formation for Quality of Communication Networks Monitoring
2018 Wave Electronics and its Application in Information and Telecommunication Systems (WECONF) Year: 2018
10. V. A. Babkin;E. P. Stroganovav Integral Quality Indicators of Modern Communication Network Functioning 2020
Systems of Signal Synchronization, Generating and Processing in Telecommunications (SYNCHROINFO) Year: 2020
11. Rong Zeng;Tianjing Liu;XuTao Yu;Zaichen Zhang Novel Channel Quality Indicator Prediction Scheme for Adaptive
Modulation and Coding in High Mobility Environments IEEE Access Year: 2019
12. Tianyu Qiu;Xiao Fu;Nicholas D. Sidiropoulos;Daniel P. Palomar MISO Channel EstimationandTrackingfromReceived
Signal Strength Feedback IEEE Transactions on Signal Processing Year: 2018.
13. Pramod Sharma, Prof. S.K Sharma, Dr. Ravindra Prakash Gupta,Performance Analysis of Wireless Sensor Network
Protocols using NS-3.17 Simulator, IJSTE Year 2015.
14. Mr. Santosh Kumar, Dr. Mohana Kumar S, Dr. J P Sridhar, Dr. Pramod Sharma, The Wireless Sensor Network of an
Adaptive Hybrid Swarm Optimization Technique for Location Privacy Using an Infrastructure-CentricMethod,NVEO
Year 2021.
15. Riya Sharma, Dr. Pramod Sharma, A Review of 5G Architecture with emphases on Security, Energy and wide
Applications. IJSRET Year 2022.

Machine Learning Based 5G Network Channel Quality Prediction

More Related Content

Similar to Machine Learning Based 5G Network Channel Quality Prediction (20)

More from IRJET Journal (20)

Recently uploaded (20)

Machine Learning Based 5G Network Channel Quality Prediction