Decentralized Methods for Cooperative Task Execution in Multi-robot Systems

Research on multi-robot systems have received a lot of interest during recent years. In this thesis, a group of methods addressing some relevant multirobot problems are presented. The first part deals with three environmental coverage problems. Chapter 1 describes the Sensor-based Random Graph (SRG) method for cooperative robot exploration. Chapter 2 describes the distributed visibility patrolling method, which can be viewed as a recurrent version of the exploration and is aimed to find optimal trajectories for a team of robots that must periodically visit a set of viewpoints in an environment with obstacles. In Chapter 3 the visibility-based pursuit-evasion method is presented, where a team of searchers must coordinate to guarantee detection of all evaders in an unknown environment while using only local information. The second part deals with localization and control problems in presence of anonymous measures. In particular Chapters 4 and 5 address the mutual localization, i.e., the estimation of the change of coordinates among the team members, using anonymous distance-bearing measures affected by false positives and false negatives. In Chapter 6 the anonymous-measurement setting is applied to the formation control field, addressing the encircling task