Energy-Efficient Path Planning of Data Collection in Multi-UAV-Assisted WSN

Unmanned Aerial Vehicles (UAVs) have been extensively studied for assisting energy supply and data collection in wireless sensor networks (WSNs). The collaboration of multiple energy-constrained UAVs is one of ways addressing the largescale coverage requirement of WSNs. In this paper, a path optimization framework of multi-UAV is proposed to effectively handle the data collection of battery-less sensor nodes. First, k-means clustering and Voronoi region algorithm are used to construct and adjust the flight boundaries of UAVs; second, paths of UAVs are optimized by Multi-agent deep deterministic policy gradient (MADDPG) algorithm to improve the energy efficiency. Simulation results show that the proposed method has significant advantages compared to existing methods. As the number of UAVs and sensors increased, the energy efficiency improved by an average of 42% compared to the Greedy strategy (GS) method and 21% compared to the multi-agent reinforcement Learning (MARL) method.