Approach to Robot Path Planning Based on Numerical Potential Field and Genetic Algorithm
Gong Jin · Journal of Tianjin University Science and Technology · 2002
A new approach to robot path planning, which can be used to plan path for any type of robot, is proposed based on numerical potential field and genetic algorithm. The approach uses geometric information of a cardinate workspace to build the distance map and the numerical potential field of a discretized workspace and uses genetic algorithm to guide robot searching on the C-space on the based heuristic function. To the local minima of a potential function defined over the C-space, a new algorithm is presented which uses genetic algorithm and alternative potential to escape the local minima. Finally, simulation experiments of path planning about a manipulator with 6 DOF′s and a two-arm cooperating robot with 12 DOF′s show the effectiveness of the proposed approach.