Containment Using Incomplete Agent Information Over a Digraph

A containment control problem deals with driving the followers in a multi-agent system (MAS) into the convex hull formed by the maneuvering leaders. The design of such a control is dependent on the measurements or exchange of complete state information, such as acceleration, velocity, and position. In this letter, we propose a containment protocol which is free from the measurements of followers’ own velocity and neighbors’ states. The leaders are maneuvering with unknown yet bounded acceleration and velocity. Only a subset of these followers are able to measure the leaders’ relative position. Since a position feedback cannot solely stabilize the second-order agents, a consensus-based observer is designed to estimate the weighted mean of the leaders’ position. The exchange of these estimates happens over a directed network (digraph). Additionally, each agent is equipped with a decentralized observer to estimate the weighted mean of leaders’ velocity. Later, a decentralized tracking controller is proposed to drive the followers’ states (position and velocity) to the respective estimates. Under the proposed containment protocol, the instability in the tracking controller of one agent does not affect other agents. Using numerical simulations, we compare the current protocol with the existing state-of-the-art to reveal better performance.