In a recent CDC submission “A closed-loop design for scalable high-order consensus,” Jonas Hansson has re-imagined the way the conventional consensus algorithm generalizes to higher-order dynamics, e.g. for formation control. This can significantly improve the dynamic behaviors in systems where feedback is limited to local, relative measurements. In particular, it alleviates the scale fragility that I described for standard high-order consensus algorithms, which implies that stability is lost if the network grows beyond a certain finite size. The benefits of Jonas’s controller come at the cost of an extra ‘hop’ in the communication network, that is, measurements need to go beyond the immediate nearest neighbor or be passed along through communication. We are excited to hear your feedback!
Emma Tegling 