Packet Reception Rate Maximization of Radio Frequency Powered Communication via Reinforcement Learning
SU Xiao-feng1，CHEN Qing-hua2
1（School of Computer Science and Technology，Zhejiang University of Technology，Hangzhou 310014，China）2（Department of Information Technology，Wenzhou Polytechnic，Wenzhou 325035，China）
Abstract：In the communication mode of RF powered，backscatter communication relies on RF signal，which leads to the problems of low transmission rate and unstable transmission；in the transmission process of wireless powered communication，it is easy to cause packet loss due to environmental interference.In order to maximize the long-term average packet reception rate，considering the limited battery capacity and the dynamic changes of the environment，this paper studies the mode selection and power allocation strategy in the wireless powered communication system assisted by backscatter under the framework of Markov decision process.The cost function is designed to represent the packet loss，and the bit error rate and packet loss rate of different communication modes are calculated.Based on this，the SARSA algorithm is used to solve the solution without prior information，and the deep Q learning method is used to solve the state space continuity problem.Finally，the simulation results show that the hybrid transmission is stable and effective in dynamic environment.In addition，the performance of online solution based on SARSA and deep Q learning is better than that of baseline solution Q-learning.
苏小枫,陈清华. 一种强化学习的射频供能通信收包率优化方法[J]. 小型微型计算机系统, 2022, 43(11): 2414-2421.
SU Xiao-feng,CHEN Qing-hua. Packet Reception Rate Maximization of Radio Frequency Powered Communication via Reinforcement Learning. Journal of Chinese Computer Systems, 2022, 43(11): 2414-2421.
2022, Vol. 43 
