Multi-Robot Exploration Inspired by Swarm Robotics using Dynamic Fuzzy Cognitive Maps

This work presents a multi-robot system (MRS) inspired by swarm robotics concepts for exploration and rescue.The exploration is done using a dynamic fuzzy cognitive map (DFCM) and fuzzy logic (FLC) to control the robots, and the results are compared for performance verification.The MRS must complete an exploration task in a semi--unknown flat environment.Thus, the MRS is not aware of the limits of the searching area, obstacles, and only knows the number of victims to be rescued as a stopping criterion.The proposed task has the purpose of simulating the rescue of victims in disasters.In the tests, the simulations used three environments to verify the global robustness in unpredictable situations, check the autonomy of the robots, and explore the areas of the performance space.The computational time between the two controller approaches is estimated using an architecture based on reactive subsumption.In general, results imply that the DFCM-based MRS approach completed the proposed tasks using less processing time and the robots had to travel less to explore a similar amount of area as seen in the FLC approach.Finally, the authors show an initial prototype, which can be used in real life experiments, and address future work.