• autonomous agents
• cooperative games
• coverage problems
• decentralized control
• Descriptor Function
• heterogeneous dynamics
• obstacle avoidance

This work of Thesis is concerned with the control of teams of autonomous agents to cooperatively perform coverage tasks. The control is set in the Framework of the Descriptor Function, which allows to deal with different scenarios, such as Deployment Problems, Effective or Persistent Coverage, in a unified fashion. The problem of full decentralization of the control law is addressed; as far as dynamic tasks are concerned, the main hurdles are the estimation of the attained coverage over time as well as a convenient deadlock escaping strategy. It is shown that if the agents are able to communicate with others within a limited range, task fulfillment is still guaranteed. Inter-agent collision avoidance is integrated with an obstacle avoidance policy that does not require an a priori knowledge of position and shape of obstacles. Without introducing further complexity, agents just need to detect their walls through proximity sensors. As regards agents’ dynamics, the state-of-the-art control laws, applied to nonlinear affine-in-control and driftless models, is adapted to a new class of vehicles with drift components, taking into account their dynamic model. Finally, Deployment Problems are framed in the context of Differential Cooperative games, for which a suboptimal solution is provided in a closed form. Throughout the work, the validity of the assertions is proved by means of formal arguments, while their effectiveness is shown in scenarios simulations aimed at comparing performances between different approaches.