Obstacle Avoidance in Dynamic Environment using Particle Swarm Optimization and Kalman Filter

Abstract

This paper presents a novel approach to tackle obstacle avoidance using the Particle Swarm Optimization (PSO) algorithm, a popular metaheuristic algorithm. One of among contributions is the introduction of the inertia rate concept, which progressively increases the emphasis on local and global optima during exploration. By incorporating this method, the performance of the PSO algorithm is significantly enhanced, resulting in improved obstacle avoidance capabilities. Comparative analysis between scenarios with and without the proposed method demonstrates a noteworthy decrease of 0.13 times in the Average collision count. To facilitate practical implementation in real-world environments, the information used in the algorithm, such as obstacle and robot positions, is subject to noise. The study conducts experiments to compare the outcomes when employing a Kalman Filter (KF) to account for this noise and when not using it.