SAPO: A UAV path optimization method through situation assessment

UAVs have become popular across various fields due to their low cost and high performance. UAV path planning, which is the basis and key of UAV motion control, is critical to achieving the autonomous intelligence of UAVs. However, offline planning’s performance is limited by the updating frequency of map information, making it challenging to optimize paths when faced with various obstacles and expanding spaces. To address this, researchers have focused on online path planning to obtain optimal paths through greedy calculations based on local details. However, local information-based planning can risk falling into local optimal solutions due to the non-convexity of path planning problems, affecting overall optimality. To solve this issue, a path optimization method based on situation assessment is proposed, which utilizes a neural network to extract critical features from sensor data. The motion state is then analyzed in conjunction with the path planning algorithm to optimize any local path that is not conducive to stable motion. By correcting the established path, the method avoids falling into the local optimal and invalid path, improving the optimization degree and feasibility of the path. Additionally, the continuity and smoothness of the trajectory are considered, and the large arc and long curve trajectory are used instead of short, broken lines to ensure the stability of the motion state. Simulation results show that the optimized path has better stability and optimality than the original path.