Flocking of Multi-agent Systems with Non-smooth Control

This paper analyzes the stability of a multi-agent system following a virtual leader. The velocity of the leader is time-varying and its acceleration is not known. The interconnection topology of position network switches based on the distances between agents and the velocity graph topology switches arbitrarily. These changes introduce discontinuities in the agent control laws. To track the leader in flocking, a neighbor-based local piecewise smooth controller is proposed for each agent. Using the control method, all agent velocities asymptotically approach the desired velocity while collisions can be avoided between agents. Some simulation results are provided to demonstrate theoretical results.