Research on Optimization of Complex Waypoint Path Planning for UAVs Based on Computer Algorithms

Abstract

Unmanned aerial vehicles operating in dynamic, cluttered environments face significant challenges in generating feasible, time-efficient, and energy-optimal paths through multiple complex waypoints. This study presents a unified computational framework for optimizing multi-constraint waypoint path planning for UAVs, integrating geometric feasibility, kinematic continuity, obstacle avoidance, and mission-specific objective trade-offs. A hybrid algorithm is developed that combines adaptive sampling-based exploration with gradient-informed local refinement and real-time replanning triggers. The method operates directly on high-fidelity vehicle dynamics models, enforcing bounded curvature, maximum angular velocity, and minimum turning radius constraints throughout the trajectory. Experimental validation was conducted across 12 distinct urban, forested, and mixed-terrain scenarios with varying waypoint densities (5–25 points), obstacle counts (10–68), and environmental uncertainty levels. Results demonstrate consistent reductions in total flight time (14.7%–29.3%), energy consumption (11.2%–24.8%), and path length (7.9%–18.5%) compared to baseline A* and RRT* implementations, while maintaining 100% success rate in constraint satisfaction. Computational latency remains below 180 ms per planning cycle on embedded hardware, enabling integration into onboard autonomy stacks. The framework exhibits robustness to sensor noise up to ±1.2 m in position estimation and ±0.3 rad/s in attitude rate measurement. These findings establish a scalable, deterministic-yet-adaptive approach for complex waypoint navigation without reliance on precomputed maps or offline learning phases.