Evolutionary approach for improvement of the operational route for mobile robots

An operational route design problem is addressed for an auto-mobile robot with four arms, which moves in a work space with a lot of examination points according to the pre-designed operational route. The operational route is designed to examine all the points in the work space with the smallest number of steps for a given landscape of the work space. Though this problem can be solved by searching a huge amount of combinations of operation points, it requires a large computation load. Furthermore, the optimal route may be too complicated. On the other hand, experienced operators can heuristically provide good operational route based on knowledge and experience. The route is usually simple but not optimal. Recognizing this fact, the paper provides a design procedure of the operation route by modifying the reference route given by experienced operators using a genetic algorithm. Computer simulation results show the validity of the proposed procedure