A Practical Genetic Algorithm for Obstacles Avoidance Robot Motion Planning

Genetic Algorithms is one of the latest AI techniques that are used in many engineering applications. Motion Planning of complex robot arms is a difficult problem that requires an efficient search and optimization algorithms. Many different approaches exist to motion planning and none of them is ideal in all situations. This paper introduces a new approach to Motion planning employing genetic algorithms. Motion planning problem is defined. Then conversion of the motion planning problem to a global optimization problem has been shown and thus solved with a genetic algorithm. Modifications of the standard genetic operators have been done to implement the proposed approach. The approach is illustrated and applied to two simple planer robot arms with 2 and 3 degrees of freedom respectively, and a more complex 4 degrees of freedom robot arm in 3D space. Cubic obstacles were considered in the motion planning environment. A program to implement this approach has been written. Finally simulation results for different motion planning problems are given.