Multi-robot System Environment To Build The Study Of The Problem

Unknown environment modelling can reflect the perception faculty and intelligent level of robots. So, research on environment modelling of mobile robots, usually named map building, receives more and more attention.In this paper, two fundamental problems for map building are introduced first, which are map representation and processing of uncertainty information. Three kinds of methods for processing of uncertainty information are described: Probabilistic theory, Dempster-Shafer theory of evidence and Fuzzy set theory. Then a method called continuum-interval based is proposed for this purpose. The upper and lower bound of continuum-interval provide the pessimistic and optimistic estimation on environment. Information degree is considered when fusion with old information.Compared with the exploration by single robot, the map got by multi-robot is more effective and accurate. On the basis of the introduction of exploration methods for single robot, exploration by multi-robot is studied which includes the control structures and cooperation methods. Two methods are addressed: dynamic region decomposition approach and behavior-selection approach. In this paper, it elaborates the basic behaviors and their implementations of each method respectively as well as how to build maps with these basic behaviors. Finally, simulated experiments and results are given.