UAV Trajectory Planning Based on Deep Residual Learning Network Optimization

Trajectory planning plays a crucial role in UAV flight to improve its efficiency and safety. Particle swarm optimization algorithm (PSO) is a popular trajectory planning method, but it can be plagued by local minima and computational complexity. In this paper, we address these issues by introducing a deep residual learning method to optimize the parameters of PSO to improve trajectory planning for UAVs. The PSO algorithm before and after the improvement is compared using a benchmark test function, which reveals the search advantages of the improved PSO algorithm. Simulation experiments are also conducted in 3D maps for comparison, proving the effectiveness of the improved PSO algorithm trajectory planning.