Multi-Objective PSO- and NPSO-based Algorithms for Robot Path Planning

Abstract

In this paper two novel Particle Swarm Optimization (PSO)-based algorithms are presented for robot path planning with respect to two objectives, the shortest and smoothest path criteria. The first algorithm is a hybrid of the PSO and the Probabilistic Roadmap (PRM) methods, in which the PSO serves as the global planner whereas the PRM performs the local planning task. The second algorithm is a combination of the New or Negative PSO (NPSO) and the PRM methods. Contrary to the basic PSO in which the best position of all particles up to the current iteration is used as a guide, the NPSO determines the most promising direction based on the negative of the worst particle position. The two objective functions are incorporated in the PSO equations, and the PSO and PRM are combined by adding good PSO particles as auxiliary nodes to the random nodes generated by the PRM. Both the PSO+PRM and NPSO+PRM algorithms are compared with the pure PRM method in path length and runtime. The results showed that the NPSO has a slight advantage over the PSO, and the generated paths are shorter and smoother than those of the PRM and are calculated in less time

Hybrid PSO-GSA robot path planning algorithm in static environments with danger zones

“…where 3 r is a uniformly distributed random variable, 1 0 3 ≤ ≤ r , and 0 3 ≥ c is a weighting factor.…”

“…The shortest path is obtained in [2] using Dijkstra's algorithm, and the PSO algorithm optimizes the path. A PSO algorithm solves in [3] the path planning problem with respect to two objectives, the shortest and the smoothest path criteria. A multi-objective PSO algorithm is used in [4] to generate trajectories for robots that are moving in environments that can contain danger sources.…”

Collision-Free Autonomous Robot Navigation in Unknown Environments Utilizing PSO for Path Planning

“…All obstacles were also presented as circles. In [18], a hybrid planning scheme was presented. In this study, PSO was used to generate the overall path, while probabilistic roadmaps (PRM) was used to refine the path around obstacles.…”

Search-evasion path planning for submarines using the Artificial Bee Colony algorithm

“…Although these stochastic approaches do not guarantee global convergence, they are able to find quasi-optimal solutions within a reasonable amount of computational time [18]. As a matter of fact, stochastic search and optimization methods have been well studied and investigated in various route planning schemes [19][20][21][22][23][24][25][26][27][28].…”