A unified communication and control approach for decentralized heading alignment in robot networks

As a critical component of collective motion coordination, decentralized heading alignment is fundamental in systems as diverse as vehicle platoons, mobile sensor networks, swarm robotics, and animal groups. Due to the intertwined interaction between nonlinear continuous-time robot/vehicle dynamics and (possibly) aperiodic discrete-time communication characteristics, a unified design of communication and control for heading alignment is extremely challenging. Existing approaches addressing decentralized heading alignment usually first perform analysis and design in the continuous-time domain and then conduct discretization-based implementation to deal with discontinuous information exchange that only happen at discrete-time instants. This two-stage strategy cannot guarantee the original designed control performance in final implementation since discretization can harm or even destabilize the heading controller designed in the continuous-time domain. Inspired by pulse based interaction in flashing fireflies and firing neurons, we propose a unified communication and control approach for collective heading alignment which addresses communication and control simultaneously and hence can retain design performance in final implementation. This approach only exchanges extremely simple and identical messages, which not only reduces processing complexity, transmission time and communication latency, but also implicitly ensures high scalability and robustness to interference in wireless communication. Experimental results on an actual swarm robotics platform are provided to demonstrate the effectiveness of proposed approach.