基于风险损失分析的多无人机巡检路径规划研究
Research on Multi-UAV Inspection Path Planning Based on Risk Loss Analysis

当前多无人机系统在执行城市巡检任务时面临巡检点的风险等级冲突性和风险损失动态性问题，为此，本文提出了一种基于风险损失分析的多无人机巡检路径规划方法。首先，该方法定义了巡检点的风险等级和损失计算方法；然后，该方法定义了一种巡检点聚类算法，根据巡检点之间的风险等级相异性和地理位置相近性对巡检点进行聚类，将不同的聚类分配给不同的无人机，避免风险等级相同的巡检点的访问冲突问题，并优化巡检路径长度；最后，该方法提出了一种基于强化学习的路径规划算法，根据实时的风险损失和巡检开销来计算奖励函数，以引导无人机的实时巡检路径规划，兼顾了动态风险规避和巡检效率。实验结果表明，本研究所提出的算法相比于现有算法具有更低的巡检风险损失以及较好的综合性能表现。
In the execution of urban inspection tasks, multi-Unmanned Aerial Vehicle (UAV) systems are currently faced with the issues of conflicting risk levels at inspection points and the dynamic nature of risk loss. To address these issues, this paper proposes a multi-UAV inspection path planning method based on risk loss analysis. Initially, this method defines the risk levels of inspection points and the approach for computing the associated losses. Subsequently, it introduces a clustering algorithm for inspection points that considers the heterogeneity of risk levels and the proximity of geographic locations, grouping inspection points into clusters. Different clusters are then allocated to different UAVs to prevent conflicts due to visiting inspection points with the same risk level and to optimize the length of the inspection path. Lastly, the method introduces a path planning algorithm based on reinforcement learning, which employs a reward function calculated from real-time risk loss and inspection costs to guide the UAVs’ real-time inspection route determination, thus considering both dynamic risk avoidance and inspection efficiency. Experimental results show that the proposed algorithm outperforms existing methods, offering lower inspection risk loss and better overall performance.