Smooth Path Planning for Mobile Robot Using Particle Swarm Optimization and Radial Basis Functions

One of the most important tasks to be performed by a mobile robot is to find a collision-free and smooth path to follow. Given a set of initial control points and using a Radial Basis Function (RBF), a method is proposed, in which is used the RBF’s property to approximate smooth functions to define a collision-free and short path. In this paper we formulate the training technique of an RBF as an optimization problem and employed Particle Swarm Optimization (PSO) to solve it. The path planning problem is equivalent to optimize the parameters of the RBF using a set of trajectory constraints based on coverage control points as input pattern, which can be seen as places where is desirable for the robot to explore. Furthermore, a combined fitness function is proposed with respect to three requirements: (i) achieve minimum mean square RBFfunction approximation error ; (ii) avoid collisions and (iii) minimize the length of the obtained path .