Research on Optimal Autonomous Navigation Inspection Path Algorithm for Distribution Network Based on Improved Simulated Annealing Method

This paper addresses the multi-constraint path planning problem encountered by vehicle-borne mobile aircraft when inspecting complex urban distribution networks. To this end, the paper proposes a UAV cooperative inspection path planning method integrating static and dynamic dual-channel data input, Monte Carlo segmentation and an improved simulated annealing algorithm. Firstly, the positions of the inspection nodes and aircraft nests are organised efficiently using a quadtree data structure, with the Euclidean distance pre-calculated. Secondly, an initial closed-loop path is constructed, and a weighted load balancer strategy is introduced. Sub-paths are then generated by combining Monte Carlo random segmentation with an improved simulated annealing algorithm, which is designed to optimise path load distribution through a dynamic heating strategy and a probability acceptance criterion. Finally, a multi-objective loss function is constructed by fusing the real-time positions of the UAV nests and traffic costs. The experiment is based on an actual distribution network comprising 414 towers. Results show that autonomous navigation planning paths generated by this method save over 20% of time compared to traditional fixed-sequence inspections, significantly improving the speed of inspection and the cooperative work efficiency of vehicle-borne mobile aircraft nests in multi-UAV distribution networks.