The effective use of multirobot systems in service applications requires the solution of several problems among which, from the Automatic Control point of view, those related to the autonomous coordination of the members of a team stand out when the collaboration among members has to bring the formation as a whole to success. The need to approach this problem on a cooperative basis stems from the fact that, even in a structured environment, the degree of a-priori knowledge of the task execution by each member of the team is intrinsically low, since the environment is shared at least with the rest of the team (which is moving, of course) if not also with other moving bodies; sometimes, as in the case of exploration tasks, the environment itself is partly or completely unknown. It is evident that, in this scenario, both navigation and motion control of the robots belonging to a team must be performed in a fundamentally different manner from that suitable to conventional contexts.

The Research Project MEMONET is aimed at the development and the experimental validation of coordinated control techniques for multirobot systems executing exploration tasks in unstructured environments. The multirobot system is organized in two groups: a first group is made of a team of exploration agents, while a second group is made of a team of service agents. The two groups perform a synergic action, in that the exploration agents build the map of the environment using the communication and localization functionalities provided by the service agents.

In this framework, the main lines characterizing the research activity subject of the Project are the following:

• cooperative methods for multirobot exploration, to accomplish synergic action of a group of autonomous vehicles. The goal is to increase efficiency of coverage of the environment to be explored considering techniques with different degrees of centralization/decentralization;
• control techniques for ad-hoc mobile networks, to exploit the collaboration among the team members by providing a service infrastructure to the mission. The goal is to guarantee the team connectivity combining the limited coverage of the single robots;
• cooperative localization algorithms, to support the dynamical generation of the reference trajectory. The goal is to exploit several available partial views of the environment to increase efficiency in reconstructing the overall state of the task so as to better drive its execution.