A hybrid model predictive control scheme for multi-agent containment and distributed sensing

This paper deals with distributed sensing over a field by means of a multi-agent control architecture. A leader/follower scheme is built up for exploring an environment by properly sensing areas of interest. By means of a control architecture based on decentralized Model Predictive Control (MPC), the leaders determine the regions to be sensed and the followers, which are only required to communicate with a subgroup of leaders, properly move in the space, constrained by the convex hull of their leaders, in order to perform the sensing task.