Velocity-consensus control for networks of multiple double-integrators

This paper studies the velocity-consensus problem of multi-agent systems with second-order agent dynamics and proposes an asynchronous distributed consensus protocol allowing for switching interaction topology, time-varying delays and intermittent information transmission. Unlike the existing works, the agreement of position-like states is not required in this paper. Velocity consensus is a prerequisite of the popular Vicsek model and can find its broad applications in congestion control for Internet, distributed task management, distributed estimation for sensor networks, etc. We first give an approach to approximating velocities based on delayed position-like information and then provide the design and analysis results for the velocity-consensus control. We show that if the union of the interaction topology across the time interval with some given length always contains a spanning tree, then the proposed asynchronous protocol will solve the velocity-consensus problem asymptotically.