Research on UAV Path Planning Strategy Based on Particle Swarm Optimization Algorithm: Performance Analysis and Optimization of Inertial Weights and Mutated Particles

Unmanned Aerial Vehicle (UAV) path planning is a critical challenge in autonomous navigation, requiring intelligent trajectory generation in complex environments that balance safety, efficiency, and adaptability. Particle Swarm Optimization (PSO), a prominent swarm intelligence algorithm, is widely used for its robust global search capabilities. However, PSOs performance heavily depends on parameter configurations, particularly the inertia weight (). This paper investigates the impact of parameter settings on PSO-based UAV path planning and proposes optimization strategies to enhance its efficacy.Simulation experiments reveal that when = 0.5, PSO exhibits strong local search capabilities and fast convergence but is prone to local optima, reducing path planning precision. Increasing enhances global search, enabling broader solution space exploration, but at the cost of slower convergence and limited optimization gains due to computational constraints. Notably, when = 1.5, the algorithm avoids local optima but shows minimal performance improvement and reduced efficiency.The study also evaluates optimization strategies, including a linearly varying inertia weight and a mutant particle strategy. The linearly varying inertia weight significantly improves optimization performance but slows convergence, making it suitable for high-accuracy scenarios with ample computational resources. The mutant particle strategy, however, achieves a better balance between optimization and convergence, ideal for practical applications requiring both efficiency and accuracy. Combining these strategies does not yield substantial improvements, suggesting that a single strategy may be more computationally efficient.This research provides valuable insights into PSO parameter configuration and optimization for UAV path planning, advancing the field of autonomous navigation technology.