Self-localization of humanoid robots using Particle Swarm Optimization algorithm

A key problem of decision making for autonomous robot is self-localization of robots toward coordinate origin. The aim of localization is finding the Cartesian coordinates and robots body direction in a global coordinate system. In this paper, we present image processing-based method for soccer humanoid robots self-localization. This method uses an inverse perspective map to convert the obtained image into taken image from top view. Also. it employs line Hough transform to modify the changing of image rotation and it defines robots location coordinates relative to the origin using particle swarm optimization algorithm. In this paper, an efficient objective function is presented to use in PSO algorithm. The most important feature of this method is use of IPM transform for deletion of perspective effects. This method relative to changing of size and form of different shapes inside of ground is resistant. Self-localization of soccer robots is one of the testing fields of this method.