A Reinforcement Learning Based on Multi-Node Charging Path Planning Method for Wireless Rechargeable Sensor Networks

Due to the limitations of sensor node battery capacity, network size, and cost, energy provision is the primary challenge in wireless sensor networks (WSNs). To solve the sensor energy limitation problem, the use of mobile chargers (MC) to supplement the energy of wireless sensor nodes has become a practical and effective approach, which is known as wireless rechargeable sensor networks (WRSNs). Compared to the one-to-one charging scheme, the one-to-many charging scheme that allows multiple sensor nodes to be charged simultaneously by a single charger can well address the charging scalability and efficiency issues. In a multi-node charging scheme, the charging locations of the MC need to be determined first before the charging path can be determined, which adds to the complexity of the charging scheme design. To address these challenges, this paper proposed a charging point determination method based on the greedy strategy and used reinforcement learning (RL) to solve the charging planning problem in WRSNs. Finally, the effectiveness of the proposed method is verified by simulation.