Nazmia Kurniawati, Y. K. Ningsih, Sofia Debi Puspa, Tri swasono adi
{"title":"Algoritma Epsilon Greedy pada Reinforcement Learning untuk Modulasi Adaptif Komunikasi Vehicle to Infrastructure (V2I)","authors":"Nazmia Kurniawati, Y. K. Ningsih, Sofia Debi Puspa, Tri swasono adi","doi":"10.26760/ELKOMIKA.V9I3.716","DOIUrl":null,"url":null,"abstract":"ABSTRAKKomunikasi Vehicle to Infrastructure (V2I) memungkinkan kendaraan dapat terhubung ke berbagai macam infrastruktur. Dengan kondisi kendaraan yang bergerak, maka kondisi lingkungan yang dilewati mempengaruhi parameter komunikasi. Implementasi modulasi adaptif pada skema V2I memperbolehkan sistem menggunakan skema modulasi yang berbeda untuk mengakomodasi perubahan kondisi lingkungan. Pada penelitian ini digunakan skema modulasi QPSK, 8PSK, dan 16-QAM dengan memanfaatkan reinforcement learning dan algoritma epsilon greedy untuk menentukan skema modulasi yang digunakan berdasarkan level AWGN. Dari hasil simulasi dengan kondisi nilai epsilon yang divariasikan dari 0.1 hingga 0.5 didapatkan bahwa semakin tinggi nilai epsilon maka semakin sering agen tidak memilih skema modulasi dengan reward tertinggi.Kata kunci: Reinforcement learning, Modulasi Adaptif, AWGN ABSTRACTVehicle to Infrastructure (V2I) communication allows vehicles to be connected to various infrastructures. Under the scenario of a moving vehicle, the environmental conditions which is passed by the vehicle will affect the communication parameters. The adaptive modulation implementation in the V2I scheme allows the system to use different modulation schemes to accommodate changing environmental conditions. In this study, the QPSK, 8PSK, and 16-QAM modulation schemes were used by utilizing reinforcement learning and the epsilon greedy algorithm to determine the modulation scheme used based on AWGN level. From the simulation results with the conditions of the epsilon value varying from 0.1 to 0.5, it is found that the higher the epsilon value, the more often the agent does not choose the modulation scheme with the highest reward.Keywords: Reinforcement learning, Adaptive Modulation, AWGN","PeriodicalId":344430,"journal":{"name":"ELKOMIKA: Jurnal Teknik Energi Elektrik, Teknik Telekomunikasi, & Teknik Elektronika","volume":"23 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"ELKOMIKA: Jurnal Teknik Energi Elektrik, Teknik Telekomunikasi, & Teknik Elektronika","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.26760/ELKOMIKA.V9I3.716","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
摘要
基础设施沟通工具使车辆能够连接到各种基础设施。随着车辆的移动状态,所经过的环境条件影响了通信参数。V2I方案的适应调整实施允许系统使用不同的调制方案来适应环境环境的变化。在这项研究中,使用q胡克、8个胡克和16个qam调制方案,利用希腊的epsilon算法和epsilon算法,根据AWGN水平确定使用的调制方案。从0。1到0。5变化值的模拟结果发现,epsilon值越高,代理选择的调制方案与最高奖励就越频繁。关键词:再生学习、适应调节、AWGN abstractlet汽车到基础设施(V2I)沟通障碍。在《a移动场景下车辆,环境的条件,这是昏倒:《communication车辆威尔影响parameters。境adaptive调制implementation V2I怀廷allows accommodate的系统要用不同的调制计划能改变环境条件。In this study,《QPSK、8PSK 16-QAM调制计划能在以前由utilizing reinforcement学习和《调制怀廷以前epsilon贪婪算法来个重大改编自AWGN级别。和条件》从《模拟results epsilon价值varying从0。1到0。5,是找到那个《epsilon价值高,经常越《探员不会选择调制怀廷with the最高奖励。重点词:恢复学习,有定型调制,AWGN
Algoritma Epsilon Greedy pada Reinforcement Learning untuk Modulasi Adaptif Komunikasi Vehicle to Infrastructure (V2I)
ABSTRAKKomunikasi Vehicle to Infrastructure (V2I) memungkinkan kendaraan dapat terhubung ke berbagai macam infrastruktur. Dengan kondisi kendaraan yang bergerak, maka kondisi lingkungan yang dilewati mempengaruhi parameter komunikasi. Implementasi modulasi adaptif pada skema V2I memperbolehkan sistem menggunakan skema modulasi yang berbeda untuk mengakomodasi perubahan kondisi lingkungan. Pada penelitian ini digunakan skema modulasi QPSK, 8PSK, dan 16-QAM dengan memanfaatkan reinforcement learning dan algoritma epsilon greedy untuk menentukan skema modulasi yang digunakan berdasarkan level AWGN. Dari hasil simulasi dengan kondisi nilai epsilon yang divariasikan dari 0.1 hingga 0.5 didapatkan bahwa semakin tinggi nilai epsilon maka semakin sering agen tidak memilih skema modulasi dengan reward tertinggi.Kata kunci: Reinforcement learning, Modulasi Adaptif, AWGN ABSTRACTVehicle to Infrastructure (V2I) communication allows vehicles to be connected to various infrastructures. Under the scenario of a moving vehicle, the environmental conditions which is passed by the vehicle will affect the communication parameters. The adaptive modulation implementation in the V2I scheme allows the system to use different modulation schemes to accommodate changing environmental conditions. In this study, the QPSK, 8PSK, and 16-QAM modulation schemes were used by utilizing reinforcement learning and the epsilon greedy algorithm to determine the modulation scheme used based on AWGN level. From the simulation results with the conditions of the epsilon value varying from 0.1 to 0.5, it is found that the higher the epsilon value, the more often the agent does not choose the modulation scheme with the highest reward.Keywords: Reinforcement learning, Adaptive Modulation, AWGN