Intelligent Dynamic Formation of the Multi-Robot Systems to Cleaning Tasks in Unstructured Environments and with a Single Perception System

Abstract

The cooperation between multiple mobile robots is a robust method to reach various objectives or expand system capabilities. A unified approach to control of robot swarm allows the perception sharing, not requiring the use of an expensive system in each agent. Thus, the information acquired by perception system of robot leader can introduce the autonomy capability in a significant number of mobile robots, sharing resources. This work discusses the use of a robot swarm to cleaning tasks with formation maintaining and obstacle avoidance, aiming to perform the job more quickly and efficiently than using only one robot or several in a disorderly way. The approach is based on a leader agent that executes the autonomous navigation through a LIDAR perception system and follower agents which are responsible for maintaining formation concerning its leader. A Fuzzy intelligent behavior commands the dynamic formation adaptation of robot swarm aiming to overcome obstacles.