The Map-explore Research Using Swarm Robots Based On Pso

The exploration of unknown area is a classic problem of the robotic research. For those areas that too dangerous or complicated for human to reach, such as disaster detection, emergency rescue, and military information-gathering, using robot to fulfill the exploration job has an important application value.Multi-robot is the hotspot of map-exploration research, and the collaborative strategies is utmost significant of multi-robot system. The traditional multi-robot coordination strategies let a few leader nodes take the responsibility of decision-making, those centralized / semi-distributed control mode is suitable for small-scale multi-robot systems. When the group size increases, the burden of computing and communications of leader nodes exponential grow.We use a swarm intelligence strategy to resolve the map exploration problem. Compared to centralized system, this non-central, non-synchronized, quasi-isomorphic and simple mechanism performs better in scalability, dynamic adaptability and robustness, and therefore more suitable for the vast, complex and unstructured environment. This subject is carried out with the following works:Firstly, we introduce the gas model to build the basic movement rules of robots, using exclusive virtual pheromone to achieve the initial formation of the groups, the spread of the movement and trends unknown area.Secondly, we design a novel"signpost– exploring"role conversion mechanism, to reduce the repeated exploration number, improve the group exploration efficiency, and enhance the dynamic environment adaptability.Thirdly, we absorb the particle swarm optimization to design the mathematical motion model based on above mechanism, and analyze/verify the suitable formula parameters.Finally, the algorithm is simulated with emluator. The performance of 30-150 group scale in different task scenarios(open, few obstacles and a number of obstacles) and different formula parameters were analyzed and assessed. Simulation results show that compared to the traditional methods, this algorithm enhance the scalability, increase the robustness and reduce the communication traffic of the system.To sum up, this subject proposed, simulated and verified a swarm robots area exploration algorithm. The simulation results show that our algorithm solved the large-scale multi-robot group collaboration problem which traditional methods are difficult to deal with.