Route Planning for Low‐Altitude UAV Using Multi‐Objective Optimization
Yong Bai, Liang Zhao · 2025
This chapter addresses the critical challenge of route planning for low-altitude unmanned aerial vehicles (UAVs) operating in complex, obstacle-rich environments below 500 meters. Traditional path planning methods often simplify multi-faceted objectives into single-criterion optimizations, failing to balance inherent trade-offs between safety, efficiency, and operational constraints. To overcome this, a multi-objective optimization framework is proposed, integrating four key normalized objectives: (1) minimizing path length, (2) maximizing obstacle avoidance, (3) ensuring altitude stability for energy efficiency, and (4) promoting trajectory smoothness to reduce mechanical stress. The UAV's kinematics and physical constraints (e.g., speed, pitch/yaw limits) are modeled using differential equations. The solution employs Multi-Objective Particle Swarm Optimization (MOPSO), which evolves a swarm of candidate paths encoded as sequences of navigation variables (segment length, pitch, yaw). MOPSO features an external archive to store non-dominated Pareto-optimal solutions, a hypergrid mechanism for diversity preservation, and adaptive mutation to prevent stagnation. Simulation tests across diverse 3D threat environments demonstrate the algorithm's efficacy. Results show that clustered obstacles necessitate trade-offs favoring safety (e.g., Case 1: path length cost 0.130, avoidance cost 0), while dispersed layouts improve path length but challenge smoothness (Case 2: smoothness cost 0.055). The approach systematically reconciles competing demands, offering robust autonomous navigation for real-world UAV applications.