Adaptive PPO-LSTM-based Particle Swarm Optimization for UAV Path Planning

Path planning for unmanned aerial vehicles (UAVs) in complex three-dimensional environments, characterized by dynamic constraints and dense obstacle distributions, remains a significant challenge. Conventional Particle Swarm Optimization (PSO) methods often struggle under such conditions due to uniform exploration behavior. To overcome the limitation, this paper proposes an Adaptive Proximal Policy Optimization and Long Short-Term Memory-based Particle Swarm Optimization (APLPSO) algorithm. APLPSO integrates Proximal Policy Optimization (PPO) and a Long Short-Term Memory (LSTM) network into the PSO framework, thereby allowing each particle to autonomously adjust its search strategy based on historical trajectories and environmental feedback. Extensive simulation studies across scenarios with varying obstacle densities demonstrate that APLPSO consistently outperforms conventional approaches in terms of path optimality, convergence rate, and robustness.