Smooth Path Planning Based on Spherical Vector Particle Swarm Optimization

Wenting Su, Zixuan Liu, Ruiyang Huang

Abstract: This paper develops a novel smooth path planning algorithm employing spherical vector particle swarm optimization to address critical challenges in UAV navigation through complex environments. The method specifically targets two persistent issues in conventional approaches: premature convergence with local optima entrapment and discontinuous generated trajectories. By establishing a bijective mapping between spherical coordinate particle positions and UAV kinematic parameters including velocity, turning rate, and climb/dive angles, the algorithm achieves comprehensive configuration space exploration for optimal cost function minimization. Subsequent path smoothing is accomplished through high-order Bézier curve control point optimization. Extensive simulations verify the proposed method’s capability to reliably produce smooth, collision-free trajectories while avoiding premature convergence, ultimately enhancing UAV path planning efficacy in obstacle-dense scenarios.

External IDs:dblp:conf/icic/SuLH25