PSO obstacle avoidance algorithm for robot in unknown environment

Abstract

PSO (particle swarm optimization) is a stochastic population based computational method that optimizes the problem iteratively, trying to improve the solution particles to get better quality of the particle called target solution. In this paper, PSO algorithm is used to find the particle best position value by searching the solution space to determine the minimum distance from the obstacle. The position of each particle is updated according to the distance and velocity equation. The LEGO NXT mobile robot which is considered to be as source is placed in any position in n-dimensional environment. And there are `n' numbers of obstacle placed in the same unknown environment. Each time, a swarm of particles are moving in the same workspace to detect and ensure if an obstacle is present there or not near to gbest location. The particles move to the global best position, following the one which is at the minimum distance from the obstacle and stop at the certain range from the obstacle. The robot then moves to the located position each time iteratively, until and unless it reaches to the target solution. Based on the position of the obstacle, the objective function to find the exact minimum distance from the obstacle is calculated. Main objective of this paper is to provide an optimized algorithm based on PSO for robot to move from source to destination by avoiding all possible obstacles. Already existing standard PSO algorithm has been modified by introducing one more objective function which is used to perform the local search based on the global search depending on the calculated gbest value by using the standard PSO algorithm. Thus we introduce a modified version of PSO algorithm called MPSO which increases the efficiency of the already existing PSO algorithm.