Path Planning for Unmanned Systems Based on Integrated Sampling Strategies and Improved PSO

B-splines and Particle Swarm Optimization algorithms are integrated for unmanned system path planning in mountainous terrains. In the early stages of the optimization search, the traditional Particle Swarm Optimization (PSO) algorithm achieves rapid convergence. However, as the process continues, it often struggles with local optima in later stages. To address this limitation, this research proposes an improved PSO algorithm that combines the Immune Algorithm (IMA) and Latin Hypercube Sampling Method. This enhancement bolsters the optimization capabilities of particles at different phases of the search by implementing an evaluation mechanism and dynamic weight adjustments. Experimental results demonstrate that, when confronting optimization challenges within complex mountainous terrains, the improved PSO algorithm (SIPSO) which is combined with IMA and Sampling Method significantly outperforms conventional PSO and Genetic Algorithm (GA) in both iteration counts and computational efficiency, showcasing a notable advancement in performance.