Research On Search Strategy Of Multi-robot Based On Discrete Event Systems

With the development of robot technology and the expansion of its application fields, Thedemand for robots is no longer confined to a single robot, but are increasingly multiple robots,The superiority of multiple robots is more prominent in fields of disaster rescue, search,stalking. Therefore, multi-robot coordinated search has become an important research topic inrobotics.An important issue in coordinated search is how to achieve coordination between thevarious robots. Each robot in the multi-robot system has autonomy, and will carry outactivities according to their own purposes, knowledge and ability in process ofProblem-solving, which often result in conflicts. Therefore coordination is needed.This paper researches discrete event-based multi-robot search strategy. Discrete eventsystems, concurrent product theory of discrete event systems, obstacle avoidance strategy andfull regional coverage by robots are combined in this paper. Multi-robot cooperative searchtask in this paper is completed with the method above. Robots are allocated reasonably toseveral local areas first. The number of robots in each local area has an upper limit, new robotis not allowed to enter the local area by the technology of mutual exclusion if its robotsnumber equals to the upper limit. Each robot within local area avoids collision with obstaclesor other robots with the obstacle avoidance strategy, which also avoid deadlocks due toresource contention. Robot path coverage algorithm based on biologically inspired neuralnetworks is used when searching to make sure the entire environment be coveraged and avoidthe inefficiency of path caused by repeated. The first left robot marks the area as “searchcompletion” when the local area search is completed. If the adjacent region is allowed to enter,the left robot move to the area, otherwise continue to look for areas not got the upper limit.All robots carry out activities like this until the search task within the entire environment iscomplete.In this article, it gives specific example for multi-robot search strategies based on discreteevent systems and analysis the case of six robots demining in ten rooms. Discrete-eventmodel of the multi-robot system, the state set, the set of events and the state transitiondiagram are also created in this paper.The initial state and all events for a hypotheticalmoment is given in this paper, the product of synthesis of discrete event systems theory cancome to the right new state for the environment is verifed later. In another way, this paper verifies the feasibility and correctness of concurrent product theory of discrete event systemsapplied to the multi-robot coordinated search system.