A hybrid algorithm of particle swarm optimization, metropolis criterion and RTS smoother for path planning of UAVs

Abstract

Particle Swarm Optimization (PSO) algorithm is a simple approach with premature convergence and stagnation prone. The loss of efficiency and sub-optimal solution occur frequently while solving path planning problem with PSO. Therefore, a method is proposed to optimize parameters which affect performance of the PSO algorithm by using Rauch–Tung–Striebel (RTS) smoother. Moreover the Metropolis Criterion is applied as acceptance policy, which can prevent the PSO algorithm from falling into local minimums in the proposed method. The RTS smoother is applied to eliminate the irregular error of the PSO updated position, and to smooth the produced path. Experimental results show the proposed method which is based on the fusion of the PSO, Metropolis Criterion and RTS performs better than the existing methods in terms of solution’s quality and robustness in the path planning problem for UAVs.