Research On Robots With Cellular Automata

Cellular automata can describe the whole complex system through simple behaviors of local cells and their interaction. In this paper, we try to do so some research work on algorithms of robot path planning and those of modular self-reconfigurable robots'locomotion planning with cellular automata.First, we describe the principle of cellular automata which is an indispensable precondition for its applying to the robot field. According to the basic principle, the cellular automata model is set to a mobile robot and its environment. When the environment is completely known, robot can get a shortest path under the rules of cellular automata. Compared with genetic algorithm and artificial potential field algorithm, the algorithm with cellular automata model is more simple , efficient and quicker. When a robot is in the environment that is partially known, a mixed algorithm including cellular automata and artificial potential field is introduced to the path planning . The simulation shows that the mixed algorithm suits the environment entirely and the robot can search a path without collision.Second, a cellular automata model is built to modular self-reconfigurable robots. Locomotion planning without obstacles and locomotion planning with obstacles are discussed, and both of them need different rules. When there are obstacles, two circumstances are considered. One is the obstacle which is higher than the robot itself, and the other is the obstacle which is lower than the robot itself. Different algorithms are applied to them and simulation proves their correctness.Third, we discuss how to deal with the broken module when a modular self-reconfigurable robot is moving ahead. An algorithm with tolerance is given and simulation proves its feasibility.