Swarm Intelligence Based Multi-robot Map Exploration Research

The problem of exploring an unknown environment belongs to thefundamental problems in mobile robotics. In many complex robot applicationscenarios whether the task is completed successfully depends on thesettlement of the map exploration problem, such as robot navigation and pathplanning, planetary exploration, disaster rescue and so on. Compared with thesingle robot system, multi-robot system is widely used in the vast, harsh andcomplex environment because of its adaptability, scalability, and highreliability. Research on it becomes a hot area of robotics research both inChina and abroad.The traditional multi-robot map exploration algorithm is only suited tothe small-scale robot groups, because of its large scale communication,presence of a central node, global communication and high redundancycoverage rate. This paper proposed a multi-robot map exploration algorithmbased on swarm intelligence. The basic idea of the algorithm is summarizedas follows: The map to be explored is divided into many subareas. Firstly, therobots diffuse to the different subarea from a small area on the edge of themap. Then the multiple robots explore the subarea in parallel until all thesubareas are explored and the allocation of subareas between different robotsfollows the particle swarm optimization model. This paper uses anisomorphic, non-central and local communication based multi-robot systemto carry out the exploration task. Compared to other map exploration strategy,the proposed algorithm performs better in efficiency, scalability and haslower redundancy coverage.The main research work in this subject is as follows: (1) The design of a multi-robot deployment algorithm based on gasmolecule model. It makes the robots diffuse to the different subarea from thesmall area on the edge of the map quickly.(2) The design of a Particle Swarm Optimization based multi-robotmap exploration algorithm. It solves two problems in the exploration stage:the subarea coverage strategy and task allocation between robots.(3) Implementation of the algorithm in simulation environment andanalysis of the experimental result. According to the different experimentalscenarios, we implement the proposed and the contrast algorithm in theplayer/stage simulator. Besides, we analyze the result to verify theperformance of the proposed algorithm.The result of the experiment shows that the deployment algorithmpresented in this paper can make the robot diffuse to the different subareaquickly. Besides, the PSO based map exploration algorithm coordinates themulti-robot completing the task well. Compared to other explorationalgorithm, the proposed algorithm performs better in scalability, adaptabilityand has lower redundancy coverage. It is suitable for the application ofmulti-robot map exploration.