智能反射面辅助的无线传感网络充电策略研究

在无线可充电传感网的研究中面临诸多挑战，如无线功率传输中易衰弱、充电规划复杂度高、内存数据易溢出导致丢包。为此，引入智能反射面（IRS）并基于强化学习设计一种无线传感网能量传输和数据收集方案IRS_MDP。首先，该方案建立反射相位偏移优化问题，计算出以任一传感器节点为充电目标节点时IRS的最优相移值，实现最大化目标节点处所接收的功率。其次，根据上述结果并结合传感器实时状态，基于强化学习设计关于充电和数据传输过程的优化充电策略实现丢包率降低。仿真结果表明，该方案相较其他方案可以更好地提升网络吞吐量和降低数据丢包率。

Research on charging strategy of intelligent reflecting surface assisted wireless sensor network

There are many challenges in the research of the wireless and rechargeable sensor network, such as wireless power transmission is prone to weakening, high complexity of charging planning and data queue overflow results in packet dropouts. Therefore, this paper led into IRS and based on reinforcement learning to design a kind of wireless sensor network energy transmission and data collection scheme named IRS_MDP. Firstly, to maximize the power received by major charging node, it presented a reflection phase deviation optimization problem and calculated the optimal phase shift value of IRS. On the basis of the above results and the real-time condition of sensors, it designed an optimal reinforcement learning charging strategy about the data transfer and charging process to reduce packet dropouts rate. Simulation results shows that this scheme can improve the network throughput and reduce packet dropouts rate better than other schemes.