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

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 within a cell is large, channel may be overloaded by signaling messages, reducing throughput or loss of data. Therefore, optimizing this signaling process is an important challenge. This thesis focuses on channel quality indicator (CQI) report irregularly transmitted from the UE to base station, and provides a mechanism to optimize 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 channel quality estimation. And proposed the utilization of Particle Swarm Optimization (PSO) in Convolutional Neural Networks (CNNs), which is one of the basic methods in deep learning. The use of PSO on the training 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 high prediction accuracy when compared to traditional methodologies.

Keywords -ChannelQualityPrediction,5GNetwork,MachineLearning,CNN,SVM

I. INTRODUCTION

Machinelearningismadeupofalgorithmsthatcanlearnfromdataandmakepredictionsbasedonwhattheyhavelearned. Thesekindsofalgorithmsmakepredictionsordecisionsbybuildingmodelsbasedontheinformationtheyaregiven,rather thanbyfollowingasetofrules.Thissetofalgorithmshasbeenusedwellinmanydifferentfields,suchascomputersecurity, bioinformatics,computervision,medicaldiagnostics,andsearchengines,tonameafew.Allofthesesystemshaveonethingin common:theycanautomaticallylookatdatabasedatatofindactionableinsightsandmakedecisionsbasedonthatdata.Mobile networksareknownforbeinghardtounderstand,anditseemslikelythatthenew5Gcommunicationsystemswillbeeven hardertounderstand.Theyneedtobeabletohandleagrowingnumberofsituationsthatcan'tbefullycommunicatedwith mobilesystemsoftoday.Someexamplesofthesescenariosaremultipledeploymentsofpowerfulpowerlines,intelligent transportationsystems,low-latencyconnections,andnetworks'company[1].Todealwiththislevelofcomplexity,weneedto comeupwithsophisticatedwaystolookat5Gdata.Inorderto makedecisions,thesemethodsneedtobeabletogather information,reducethenumberofpeopleneededtorunthesecommunicationnetworks,cutdownontheamountofworkthat comeswithmanagingnetworks,andpredicthowusersandnetworkswillactinthefuture.

Channel Quality Indicator-

TheChannelQualityIndex,whichisalsowrittenasCQI,isameasureofhowwellinformationcanbesentoverawireless channel.ACQIcanbeavalueormanyvalues,anditstandsforametricthatmeasureshowgoodacertainchannelis.Mostof thetime,aCQIwithahighchargemeansthatthechannelalsohasahighcharge,andviceversa.Useperformanceindicators likethesignal-to-noiseratio(SNR),thesignal-to-interferenceplusnoiseratio(SINR),thesignal-to-noiseplusdistortionratio (SNDR),etc.tofigureouttheChannelQualityIndex(CQI).Byfiguringouttheseorothervaluesforachannel,youcanthenuse thosevaluesto figure outtheCQIfor thatchannel[2].TheCQIofthechannel canbeaffected bythetypeoftransmission (modulation)usedbythecommunicationsystem.Forexample,acommunicationcompanythatusesmulti-inputdistribution code(CDMA)canuseawiderrangeofCQIsthanonethatusesorthogonaldivisionmultiplexing(OFDM).Inmorecomplicated communicationsystems,likethosethatusemulti-channelinput(MIMO)orspacecoding,theCQImayalsodependonthetype ofreceiverbeingused.ThingsthatcanbetakenintoaccountinCQIincludethefailuretocarryoutthedemonstration,Doppler shift,theevaluationoftheinformationchannel,interference,andsoon.

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

II. RELATED WORK

Sihem Bakri et.al. 2020 [1] Channelqualityfeedbackiscrucialfortheoperationof4Gand5Gradionetworks,asitallows controllingUserEquipment(UE)connectivity,transmissionscheduling,andthemodulationandrateofthedatatransmitted overthewirelesslink.However,whensuchfeedbackisfrequentandthenumberofUEsinacellislarge,thechannelmaybe overloaded by signaling messages, resulting in lower throughput and data loss. Optimizing this signaling process thus representsakeychallenge.Inthispaper,wefocusonChannelQualityIndicator(CQI)reportsthatareperiodicallysentfroma UEtothebasestation,andproposemechanismstooptimizethereportingprocesswiththeaimofreducingsignalingoverhead andavoidingtheassociatedchanneloverloads,particularlywhenchannelconditionsarestable.Tothisend,weapplymachine learningmechanismstopredictchannelstability,whichcanbeusedtodecideiftheCQIofaUEisnecessarytobereported,and inturntocontrolthereportingfrequency.Westudytwomachinelearningmodelsforthispurpose,namelySupportVector Machines(SVM)andNeuralNetworks(NN).Simulationresultsshowthatbothprovideahighpredictionaccuracy,withNN consistentlyoutperformingSVMinoursettings,especiallyasCQIreportingfrequencyreducesobtainedthebestASEusingthis method.

Lubov Berkman et.al (2019)[2] thechangeinchannelcapabilityappearsasachangeinanumberofobjectives,whichare proposed todeterminethe valueofstate-of-the-artcontrol technology.Itis recommendedtousethe gradient prediction methodtopredictthestateofthechannel.Analyzetheparametersthatcharacterizetheseparatechannelstate.Thecontrol elementneedstoconsiderthecharacteristicsofthecontrolchannelandputinplacetoimprovetheefficiencyofthecontrol. Considerhowtoevaluatethequalityofacommunicationchannel.Wedefinethealgorithmthatisproposedtobeusedona networkwithpacketpacketsduringtheprocessingofaccesstolimitthenetworkbandwidth.Accordingtothegeneraland partialdetailsoftheimprovement,considerthepossibilityofchoosingthebestmethod.Analyzethemeasurementofquality andservicestandards.Itisveryconvenienttouseacontinuousmeasurementmethodandaclearoutputmethodtomeasure theload.Anetworkexchangespackageswhenmeasuringservicequalityindicators(numberofreports,expectedtimetostart serviceandτ).π)Itisveryconvenienttousethemethodofdirectlycountingthenumberofreports.

V. A. Babkin et.al (2019)[3] Inordertoensurethequalityoftrafficflowinacommunicationnetwork,itisnecessarytoensure thevalueofthequalityindexwithinanacceptabletimeframe.Oneoftheseindicatorsisthetraffictransmissionratereported inthetrafficdataprocessing.Bycheckingwhethertheuser'strafficmanagementfilematchestheconfigurationfilespecifiedin theconfigurationfile,thequalitycontrolvaluecanbekeptwithinasinglevalue.ofthetraffic,thusmaintainingthequalityof theuser’straffic.

Hesham M. Elmaghraby et al. (2018)[4]thispapersolvestheproblemofchanneldistributionforfemtocellsthatsharethe commonuseofmacrocells.Theprogramproblemofthefemtobasecamp(FBS)ispresentedintheformofaRestlessArmed Rogue(RMAB)system.Ourgoalistoselectabranch/channelthatoptimizetheamountofexpectedreductionrewardoveran indefiniteperiodoftime,whileminimizingtheinterferencecausedbycelldivisionchanneldistribution.Insteadofdirectly monitoringtheactualchannelquality,weuseacellularuserfeedbackcalledtheChannelQualityIndex(CQI).Ingeneral,the RMABproblemisaPSPACEproblem.InordertoestimatetheavailablechannelsintheFBS,weproposeanindexingstrategy withlowinferencedifficulty,calledtheWitLeaverageindex.Findingaclosedchannelreservationsolutionoftenmeansthat thereareclosedchannelreservationsthathaveanactiveprogrambutarebasedonpartialchannelinformationintheCQI.We alsohighlightthebenefitsofareferralpolicyoverashort-sightedpolicy.

III .PROPOSED SYSTEM

5Gnetworkscangiveusersabetterexperiencebecausetheyhavemorecapacityorbettermanagement,buttheyalsoneed moreaccuratechannelpredictionsthanoldermobilenetworksdid.Inthisthesis,amachinelearningmethod,includingthe CNNandSVMalgorithms,wassuggestedasawaytopredicttheChannelQualityIndex(CQI).Reflection,diffraction,andsignal scatteringarethethreethingsthat,inatypicalcellularcommunicationscenario,WhencomparedtoLTE,thephysicallayer resourcesavailableina5Gnetworkaremoreplentifulbutalsomoredifficult.Asaresult,algorithmsforschedulingthatare moreflexibleanddependable,inadditiontoCQIvaluesthataremoreaccurate,areofutmostsignificancefortheadvancement ofNR.networks.CQIcreationandreportinghavetraditionallybeencarriedoutbyfavortheideaofdelayingthetimetable, whichwillinevitablyleadtoadecreaseinthesystem'soverallperformance.UtilizingtoolsofpredictionsuchasTheCQI's accuracycanbeimprovedwiththehelpofdeeplearningtechniques.whichisfinallysomethingthathelpsouttheNRsystem.

0+/8A.The Prediction ofCQIThrough the UseofDeep LearningAlgorithmsAt the moment, weareconcentratingonthe widebandCQI,whichisapositiveinteger.between1and15.ConsideringthattheCQIismerelyadiscretevaluewhileUEsare continuous,theItisnotpossibletomerelytrainbecauseconditionsareconstantlyshifting.acomponentthatcanpredictthe actionsofanyuserbasedsolelyontheirhistoryimportanceoftheCQI.Intheeventthatauserisgoinginthedirectionofthe

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

basestationatwhereasatothertimesitismovingawayfromtheBS,itismovingawayfromtheBSthistime.Becauseofthis, themodulewon'tknowhowtorespondtotheuser'sactions.Afterward,allitdoestolearnishowtomimictheconductithas alreadyseen.Thisbeingsaid,inthismanner,thetrainingmodulewillonlyrequireafewstepstocomplete,andtheresult indicatesthattherewasanapparentdelayintheCQIresultwhencomparedtotheactualvaluethathasbeenreported.In additiontothis,theactionstakenbyusersarvary,hencetheBSoughttohavedifferentmodelsforeachsinglevariation.users whoconnectedthemselvestoit.

Figure1 Illustration

ofCQIprediction

module.

Atthephysicallayerofa5Gnetwork,therearenotonlymoreandbetterresourcesthaninanLTEnetwork,buttherearealso moreofthem.Becauseofthis,itisveryimportantto improveNR networksby makingtheirscheduling algorithmsmore flexibleandreliableandbygivingthemmoreaccurateCQIvalues.OverthecourseofCQI'shistory,itsproductionandreporting havebeenknowntocausescheduledelays,whichinturnhaveledtoadropinsystemperformance.Wecanimprovethe accuracyoftheCQIbyusingpredictionmethodslikedeeplearningalgorithms,whichisgoodfortheNRsysteminthelongrun.

Adownlink scheduler, whichisalsocalleda MACScheduler,is part oftheNR system'smediumaccess(MAC)layer. This scheduler'sjobistogettheuser'spersonalscheduleinformationwhentheuserconnectstothebasestation.Thisinformation includestheCQIaswellasQualityofServicemessagesandbufferstatusreports(BSRs)sentbytheRadioLinkControl(RLC) layer(QoS).TheschedulerwillthenpickausertorepresenteachRBbasedontheinformationthathasalreadybeengiven.The valueoftheCQIisusedtofigureouttheMCS,whichthengivesinformationabouttheTransportBlockSize.WhenRBsaregiven outtousers,thisiswhathappens(TBS).Therearelimitsonhowmuchdatacanbesentduringthistime,whicharesetbyboth theMCSandtheTBS.UsershavetogivetheCQIateverytimeintervalthathasalreadybeenset.Thebasestationhadtosend theCQIbeforeitcouldgetthefeedback,sotherewillbeadelaybetweenwhenitasksfortheCQIfeedbackandwhenitgets it[5-7].

Atthephysicallayerofa5Gnetwork,therearenotonlymoreandbetterresourcesthaninanLTEnetwork,buttherearealso moreofthem.Becauseofthis,itisveryimportantto improveNR networksby makingtheirschedulingalgorithmsmore flexibleandreliableandbygivingthemmoreaccurateCQIvalues.OverthecourseofCQI'shistory,itsproductionandreporting havebeenknowntocausescheduledelays,whichinturnledtoadropinsystemperformance.Wecanimprovetheaccuracyof theCQIbyusingpredictionmethodslikedeeplearningalgorithms.ThisisgoodfortheNRsysteminthelongrun.

A downlink scheduler, also called a MAC Scheduler, is part of the NR system's medium access (MAC) layer. When a user connectstothebasestation,itisthejobofthisschedulertogettheuser'spersonalscheduleinformation.

ThisinformationincludestheCQIaswellasQualityofServicemessagesandbufferstatusreports(BSRs)thataresentbythe RadioLinkControl(RLC)layer(QoS).TheschedulerwillthenchooseausertorepresenteachRB,takingintoaccountthe informationthatwasgivenbefore.ThevalueoftheCQIisusedtocalculatetheMCS,which,inturn,tellsusabouttheTransport BlockSize.

WhenRBsaregiventousers,thisiswhathappens(TBS)[8].TheMCSandtheTBSbothputlimitsonhowmuchdatacanbe sentatthistime.UsersarerequiredtogivetheCQIateverypredeterminedtimeinterval.ThebasestationhadtosendtheCQI beforeitcouldgetthefeedback,sotherewillbeadelaybetweenwhenitasksfortheCQIfeedbackandwhenitfinallygetsit.

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

Figure. 2 FlowDiagram

IV SYSTEM DESCRIPTION

Using machine learning to make a system that can predict the future To make a prediction system, a machine learning algorithmgoesthroughthenexttwosteps.

a) Training phase: Seventypercentofthefeaturevectorsandtheirlabelsareusedtotraintheclassifier(whichreflectthe actualclasses).Duringthetrainingstepofmachinelearning,afunctionthatmapsinputs(called"featurevectors")tooutputsis made(labels).Then,thisfunctionisusedtoputnewvectorsintogroups.Atthispoint,theNNalgorithmcanhandlebothlinear andnonlinearfunctions,whiletheSVMmethodonlylearnslinearfunctions.

b) Test and validation phase: Inthisstep,weusetheremainingfeaturevectors(30percent).Itinvolvescomparingthe predictedclassesforthesevectorswiththelabelsthathavealreadybeengiventothem.

Particle Swarm Optimization (PSO)- Afterthat,thismethodcanbeusedtoimprovethewayCQIdatamessagesaresent, whichwillleadtolesssignallingoverheadinthelongrun.AndshowedhowParticleSwarmOptimization(PSO),whichisoneof themostimportantdeeplearningtechniques,canbeusedinConvolutionalNeuralNetworks(CNNs).Duringthetrainingphase, PSOisusedtoimprovetheaccuracyoftherecognitionprocessbymakingtheresultsoftheCNN-generatedsolutionvectorsas goodaspossible.

ImagineanetworkwithNnodesofuserequipmentandonebasestation.Allofthenodescantalktoeachother(BS).Aspartof theseconversations,theChannelQualityIndex(CQI)ofthecommunicationfrequencybandsisgivensothattheconditionsof thosefrequencybandscanbelookedat.Thismakesiteasierforpeopletotalktoeachotherinawaythatworksbetter.Because ofthis,thesignal-to-noiseratio(SNR)foreachsubcarrierisagoodchoiceofCQI,anditwillbeusedassuchfortherestofthis investigation.Inreality,theCQIiseithera4-bitvalue(for5G)ora5-bitnumber,anditencodesthechannelgainsaswellasthe modulationandcodingschemebeingused(MCS).Subbandsaresmaller,morespecificpartsofeachfrequencybandwidththat aremadeasitisdividedfurther.Afterthat,eachsub-bandisdividedintophysicalresourceblocks(PRB),whicharemadeupof sub-carriersintheend(SC)

Table 1

-NetworkParameters

Parameters Value

Numberofnodes 150

Arealength(m) 400

Areawidth(M) 400

Expectedsensorsreadings 30

Sizeofdatabuckets(bytes) 400

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

5G Dataset with Channel and Context Metrics-Anexampleofa5GtracefilefromalargemobilenetworkproviderinIreland, whichispartoftheUnitedKingdom.Irelandishometotheheadquartersofthecompanythatrunsthisservice.Thedataset wasmadebyputtingtogethertwodifferentapplicationpatternsandtwodifferentmobilitypatterns thestaticpatternandthe vehiclepattern(videostreamingandfiledownload).Duringthedatacollectionprocess,thefollowingarethemostimportant performanceindicatorsoftheclient-sidecellularnetwork:(KPIs).Someofthekey performanceindicators(KPIs)thatare includedherearethroughput,indicatorsaboutcells,metricsaboutcontext,andmetricsaboutchannels.Awell-knownAndroid programmecalledG-NetTrackProwasusedtotakethesemeasurements.Itwasmadefornetworkmonitoringanddoesn't requirethedevicetoberootedinordertowork.Asfarasweknow,thisisthefirsttimethatinformationaboutthethroughput, channel,andcontextof5Gnetworkshasbeenmadepublic[9].

Figure 3.dataset

Bothareal-time5GproductionnetworkdatasetandaMATLABmodellingframeworkforlarge-scalemulti-cell5Gnetworks. Nowthatthe5G/mmwavemoduleforthens-3mmwavenetworksimulatorisavailable,wewillbeabletolearnmoreabout howadaptiveclientsin5Gmulti-cellwirelesssituationscometotheirconclusions.

Themaingoalofourframeworkistogiveendusersmoreinformation,suchasavarietyofmetricsthatareonlyrelevantto userswhoareconnectedtothesamecell10-12].Thiswillgiveendusersaccesstoinformationtheycouldn'tgetanyotherway about the environment of the base station (eNodeB or eNB) and the scheduling principle. We make it possible for other academicstolookintothisinteractionbylettingthemuseourtechnologytomaketheirownfakedatasets.

V SIMULATION RESULTS

Whenjudgingthequalityofthecommunicationmodelasawhole,itisimportanttotakeintoaccountthesignal-to-noiseratio. TheModulationSchemesandtheCommonQualityIndex.Itisveryimportanttounderstandhoweachofthesepartsfitsinto thewhole.Thecorrelationqualityindex(CQI)andthesignal-to-noiseratio(SNR)arerelated,accordingtotheresultsofthis researchproject(CQI).Itwasdecidedthatthisdealwouldgiveabetterdataratethanotherpartnershipsthatwerealreadyin place.Whenjudgingtheoverallqualityofacommunicationmodel,itisimportanttolookattheSignal-to-NoiseRatio(SNR),the ChannelQualityIndex(CQI),andtheModulationSchemes.Itisveryimportanttounderstandhoweachofthesepartsfitsinto thewhole.Thecorrelationqualityindex(CQI)andthesignal-to-noiseratio(SNR)arerelated,accordingtotheresultsofthis researchproject(CQI).Itwasdecidedthatthisdealwouldgiveabetterdataratethanotherpartnershipsthatwerealreadyin place.

value:

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

Figure.4 parametersinitialization

Fig5windowshowingtheparametersinitialization,theinitialparametersgivennumberofnodes,area,widths,packetsrates andexpectedsensorreading.

Figure 5 nodedeployedinthenetwork

2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified

Fig

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

6showingthenodedeployedinthenetwork,inthissimulation150nodesdeployedinthenetwork

Figure.6 networkcreationandnodedeploymentofnodes

Thefigure7depictstheinitializationofparametersinthisnetwork,whichhas150nodesandalengthandwidthof400m.and sizeofthedatabucket400.

Figure.7 checkqualityofnodeinthenetwork

Afterthenetworkhasbeensetup,allofthenodesarespreadoutrandomlyacrossthenetwork.Fromthere,packetsaresent fromonenodetothenext.Alloverthebluenode,youcanseewherethedatanodesare.Thisnode'sjobistomakeitpossible forinformationtomovefromonenodetoanother.Figure8showsusthis.

2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified

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

Figure. 8 checkqualityofnodeinthenetwork

Somenodesmaybeweakandunabletosendpacketsfromonenodetoanotherwhileothersaresendingpacketsfromone nodetoanother.Machinelearningwasusedtocheckthequalityofthenode,andtheweaknodeinthenetworkisshownbythe rednode.Figure9showsthechannelqualityindicatorforthisnetwork,whichwasbasedonthechannelqualitydataset.

Figure.9 weaknodes

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

Fig10showingtheweaknodesinthenetworkthosenodenotfullfilltherequirementofthechannelqualityconsiderasaweak nodeinthenetwork.

Figure.10 reconstructnodeinthenetwork

Afterapplyingamachinelearningalgorithm,thegreennoderepresentsthereconstructednodeinthenetwork,showinginfig 11

Fig.11-performanceofnetwork

Fig.12showingtheperformanceofthenetworkintermsofPSNR,BER,MSE,accuracy.

CQI Value Ranges=Onascalefrom0to30,where30meansthechannelisthebestand0meansitistheworst,30isthebest ratingthatcanbegiventothechannel.Thesizeofthetransportblocksthatthenetworkusestosenddatavariesandisbased onwhatis reportedtotheEU.Whenuserequipment(UE)transmitsahighCQItothenetwork,thenetwork responds by sendinglargerblocksizes.Whentheoppositeistrue,smallerblocksizesareusedtosendinformation.

2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified

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

EventhoughtheusersaysthattheCQIislow,itisstillpossiblethatthenetworkissendingalotofdata.IftheUEhasaCRC error,itislikelythatitwon'tbeabletofigureoutwhattheinformationis.Becauseofthis,thenetworkwillhavetosendit again,whichisawasteoftheradioresourcesthatareavailable.

WhatshouldauserdoiftheactualchannelqualityisloweventhoughtheUEclaimstohaveahighCQI?Inthiscase,ifthe networksendsabigtransportblocksize,it'smorelikelythattheUEwon'tbeabletodecodeit,whichwouldcauseaCRCerror ontheUEsideofthecommunication.Thiswouldmeanthatthenetworkwouldhavetosenditagain,whichwouldwasteradio resources.Inthiscase,theCQIvaluewilltellthenetworkthateachtransportblockneedstocarryalotofdata.

Fig. 12-PerformanceGraph

Training Results



numbatches=0.0200  epoch1/1  Elapsedtimeis0.157340seconds.  Trainingerror=100%  TIME2= 0.426  Testingerror=150Nodes  TIME1=5.4759

Performance Evaluation Model-BitErrorRate(BER):TheBitErrorRate(BER)isawaytofigureouthowmanybiterrors happeninacertainamountoftime.Dividethetotalnumberofbiterrorsthathappenedduringthistimeperiodbythetotal numberofbitsthatweresent.Thisiscalledabiterrorrate(BER)(BER).BERisalmostalwaysgivenasapercentagelowerthan unitswhenmeasuringperformance.Tofigureouthowlikelyitisthatabitwillgowrong,youmustfirstfigureoutthepredicted biterrorrate.Thebiterrorrateisameasurementthatcanstandinforthechancethatabiterrorwillhappen.Thisestimateis correctfortimesthatarelongerthanonesecondandforerrorsthatinvolvemorethanonebit.Theformulacanbeusedto showthattheBERisafunctionofEb/N0forbothQPSKandAWGNmodulation.

BER= erfc( / )

Thereareseveralwaystofigureouthowgoodapicturecompressionis.Twoofthesewaysarethemeansquareerror(MSE) andthepeaksignal-to-noiseratio(PSNR).TheMSEstatisticshowsthetotalsquarederrorbetweenthecompressedimageand theoriginalimage,whilethePSNRstatisticshowstheerrorinitsworstform.Asthetotalnumberoferrorsgoesup,theMSE valuewillstarttogodown.

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

Peak signal-to-noise ratio (PSNR) -Itisthesignal-to-noiseratiothatshowshowwellasignalcanbeshownatitshighest possiblevalue(power)comparedtohowmuchnoiseaffectsitsaccuracy(PSNR).Thispercentageisshownasanumber.

PSNR = 10log10( ) 10.log10 ( ) 20.log10(MAXI)-10.log10(MSE)

TheMAXIcommandshowsthemaximumvaluethatcanbeusedforagivenpixelinanimage.Using8bitspersampletoshow thepixels,thisvalueisthesameasthenumber255.WhensamplesareencodedwithlinearPCMandBbitspersample,MAXIis oftenthesameas2B1.

Mean squared error (MSE) -MSEorMSDisawaytomeasuretheaveragesquareofanestimator'smistakesortheaverage squareofthedifferencebetweenwhatwasestimatedandwhatwasmeasured.Thisnumberisalsocalledthe"meansquared error,"the"meansquareddeviation,"andothernames.

MSE =

MSE} = meansquarederror

{n} = numberofdatapoints

Y{i} = observedvalues

{Y}{i} = predictedvalues

Proposed System

Table 2- Comparisonresultwithexitingwork

Technique

Optimization Technique Accuracy (%)

ConvolutionNeuralNetwork ParticleSwarmOptimization(PSO) 99.41 SupportVectorMachine 95.21

Existing System SupportVectorMachine 92.86

2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified

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

Table 3-Performanceofproposedalgorithm

Technique BER Mean Square Error Peak Signal-To-Noise Ratio True Positive Rate

Proposed System Convolution NeuralNetwork 0.41 1.05 47.8 0.12

V CONCLUSION

TheroutinesendingofCQIreports,whichgiveinformationaboutthequalityofthechannelsin4Gand5Gmobilenetworks, addstotheamountofsignallingthatneedstobedonetokeepthenetworksrunning.TheCQIreportstellushowgoodthe channelsin4Gand5Gmobilenetworksare.Welookedintoalotofdifferentwaystolowerthetotalcost.Wethinkthatifwe thinkabouthowstablethechannelis,wewon'thavetosendCQIsignalswhenit'snotnecessarytodoso.Toputitanotherway, wewillsendoutfewerCQIreportswhenthevalueoftheCQIdoesn'tchangemuchovertime.Thismeansthatthechannelis workingasitshould.Todothis,weputalotoftimeandeffortintomakingmachinelearning-basedstrategiesthatonlyneed CQIdataasaninput.Thegoalofthisexercisewastofigureouthowtomakeaccuratepredictionsabouthowthechannelwould behave.So,thewayourmechanismsworkisinlinewiththestandardsanddoesn'tneedanycross-layerorotheroutside information,likewheretheusersareorhowtheymovearoundtheenvironment.

Inthisparticularsituation,welookedathowwellSupportVectorMachines(SVM)andConvolutionNeuralNetworks(CNN) couldpredicttheresults.Wealsolookedintowhetherthereisalink betweenhowaccurateourpredictionsareandhow quicklywegetnewinformation.TheresultsofourtestsshowedthatneuralnetworksalwaysdidbetterthanSVMsinevery situationwetestedthemin.Duringthisinvestigation,wespentmostofourtimetryingtofigureouthowwelltheMLsystems wewerelookingatcouldpredictthefuture.Thenextstepistostartamorein-depthstudyofhowtheproposedmethodand processesaffectthe5Gnetworkslicemanagementarchitecturethatweproposedinourearlierstudy.Thenextstepinthe processwillbetodothis.Atthemoment,ourmaingoalistofigureouthowtoimprovesignalingandwhateffectourideaswill haveonhowavailableresourcesareusedandhowwell5Gnetworks

References

1. SihemBakri∗,MahaBouaziz∗,PantelisA.Frangoudis‡,andAdlenKsentini∗Channelstabilitypredictiontooptimize signalingoverheadin5GnetworksusingmachinelearningSihemBakri∗,MahaBouaziz∗,PantelisA.Frangoudis‡,and AdlenKsentini∗978-1-7281-5089-5/20/$31.00©2020IEEE

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

2. LubovBerkman;OlhaTkachenko;LarysaKriuchkova;OksanaVarfolomeevaDeterminationofCriteriaforChoosingthe BestWaysandIndicatorsofServiceQualityinInfocommunicationNetworks20193rdInternationalConferenceon AdvancedInformationandCommunicationsTechnologies(AICT)Year:2019

3. V.A.Babkin;E.P.StroganovaIntegralQualityControlofServiceProfileofTraffic2019SystemsofSignalsGenerating andProcessingintheFieldofonBoardCommunicationsYear:2019

4. HaifengGuoModelPredictiveControlonDual-ChannelClosed-LoopSupplyChainModelandItsBullwhipEffect2018 EighthInternationalConferenceonInstrumentation&Measurement,Computer,CommunicationandControl(IMCCC) Year:2018

5. Pengfei Li;Yun-Bo Zhao;Yu Kang Integrated Channel-Aware Scheduling and Packet-Based Predictive Control for WirelessCloudControlSystemsIEEETransactionsonCyberneticsYear:2020

6. Chi-ChuanHo;Bo-HongHuang;Meng-TingWu;Tin-YuWuOptimizedBaseStationAllocationforPlatooningVehicles Underway by Using Deep Learning Algorithm Based on 5G-V2X 2019 IEEE 8th Global Conference on Consumer Electronics(GCCE)Year:2019

7. SahrishKhanTayyaba;HasanAliKhattak;AhmadAlmogren;MunamAliShah;IkramUdDin;IbrahimAlkhalifa;Mohsen Guizani5GVehicularNetworkResourceManagementforImprovingRadioAccessThroughMachineLearningIEEE AccessYear:2020

8. Joel Shodamola;Usama Masood;Marvin Manalastas;Ali Imran A Machine Learning based Framework for KPI MaximizationinEmergingNetworksusingMobilityParameters2020IEEEInternationalBlackSeaConferenceon CommunicationsandNetworking(BlackSeaCom)Year:2020

9. V.A.Babkin;E.P.StroganovaPrinciplesofIndicatorsFormationforQualityofCommunicationNetworksMonitoring 2018WaveElectronicsanditsApplicationinInformationandTelecommunicationSystems(WECONF)Year:2018

10. V. A. Babkin;E. P. Stroganovav Integral Quality Indicators of Modern Communication Network Functioning 2020 SystemsofSignalSynchronization,GeneratingandProcessinginTelecommunications(SYNCHROINFO)Year:2020

11. RongZeng;TianjingLiu;XuTaoYu;ZaichenZhangNovelChannelQualityIndicatorPredictionSchemeforAdaptive ModulationandCodinginHighMobilityEnvironmentsIEEEAccessYear:2019

12. TianyuQiu;XiaoFu;NicholasD.Sidiropoulos;DanielP.PalomarMISOChannelEstimationandTrackingfromReceived SignalStrengthFeedbackIEEETransactionsonSignalProcessingYear:2018.

13. PramodSharma,Prof.S.KSharma,Dr.RavindraPrakashGupta,PerformanceAnalysisofWirelessSensorNetwork ProtocolsusingNS-3.17Simulator,IJSTEYear2015.

14. Mr.SantoshKumar,Dr.MohanaKumarS,Dr.JPSridhar,Dr.PramodSharma, TheWirelessSensorNetworkofan AdaptiveHybridSwarmOptimizationTechniqueforLocationPrivacyUsinganInfrastructure-CentricMethod,NVEO Year2021.

15. RiyaSharma,Dr.PramodSharma,AReviewof5GArchitecturewithemphasesonSecurity,Energyandwide Applications.IJSRETYear2022.