Research on Intelligent Flight Path Planning and Simulation for UAV Inspections in Complex Grid Environments

Unmanned aerial vehicles (UAVs) have become important equipment for overhead power grid inspection, yet general-purpose path planning methods are inefficient in large mountainous inspection areas and produce jagged paths that violate UAV turning limits. This paper presents an integrated pre-mission path planning framework for multi-target power grid inspection. First, SRTM DEM data are rasterized to model the terrain of the inspection area. Second, a sparse A* algorithm with flight constraints (heading angle, pitch angle, minimum altitude, and minimum step size) is improved by solution space pruning and a node generation search, and Theta*-based line-of-sight smoothing removes jagged grid paths. Third, an adaptive genetic algorithm with dynamic crossover and mutation probabilities solves the multi-target traversal problem by jointly considering flight distance and turning angle. In simulations on three terrain scenarios, the proposed method reduces the number of expanded nodes by about 65%, computation time by 37–39%, and path length by 4–5% compared with the standard sparse A*. Experiments with 10–30 randomly generated target points on real terrain data from western Beijing provide an initial indication of scalability.