Multi-robot Coordination Technology Research

The article makes researches on arithmetic for multi-robot avoid block and multi-robot formation control which are two typical multi-robot problems.As to multi-robot avoid block,divide robot's behaviors into five which are called move-to-goal, avoid-static-block,fall-wall,avoid-robots, rand-move. Robots get surrounding partial information by sensors and get overall information by robot observing target. In ordor to avoid the static block, artificial power theory is used. To avoid other robots,it is better to adopt traffic regulation method. When robot's surrounding is narrow or robots are more but cause robot motionless, rand disturbance can probably make robot rand move. While rand disturbance takes no effect,then the robot will adopt fall-wall behavior to get away from deadlock.As to multi-robot formation control, first ,divide robot's behavio- rs into five which are called move-to-goal, keep-formation, avoid-static-block,avoid-robot, rand-move; then the experience function gives each behavior's control parameter that is used in the combination of behaviors. Similar to multi-robot avoid block, for the merit and shortcoming that both partial information and overall information have, decision analysis combines two kinds of these methods to make a decision. Adopt leader-folower method to keep formation, the main mission of leader is leading all robots to target, followers regard leader as reference to keep own formation position in time. For better formation control, leader is changeable according to smart strategy . Also use rand-move and fall-wall behaviors to make the robot be able to get away from deadlock.For environment that robot faces is usually uncertain and robot state is diverse,robot may have several choose. So the designers must completely consider every situlation that robot may face.However,it is not realistic and the efficiency is very low. Let robots study independently so to adapt the environment is a kind of development trend. The genetic algorithm is a very effective optimize method, use the genetic algorithm to optimize the formation control parameters which is used in the combination of behaviors.In ordor to verified above arithmetic, carry out many simulation experiments under the software matlab 6.5 environment, the results shows the feasibility of the arithmetic.