Distributed control algorithm and experiment of multi-agents in a non-rectangular bounded space

In this paper, we consider coordinated motion and cooperative control of multi-agents in a non-rectangular bounded space, and present a velocity consensus algorithm for the agents with double-integrator dynamics. The traditional consensus algorithm for bounded space is only applied to rectangular bouncing boundaries, not suitable for non-rectangular space. Therefore, we introduce the concept of the mirrored velocity that not only can convert the discontinuous real velocity into the continuous mirrored velocity, but also can expand a bounded space into an infinite space, and eventually the velocity of multi-agents asymptotically converge to the same values, respectively. Finally, the effectiveness of the proposed consensus algorithm is examined by numerical simulations. Moreover, we pave the way from theoretical design to practical implementation by performing an experiment in a real multi-robot system.