DESIGN OF ARBITRARY ORDER DYNAMICAL OUTPUT FEEDBACK PROTOCOLS FOR MULTI-AGENT SYSTEMS OVER DIRECTED COMMUNICATION GRAPHS
Multi-agent Systems, Linear Systems, Control, Linear Matrix Inequalities
This work proposes new conditions for controlling homogeneous multi-agent systems using dynamic output feedback protocols. The agents are described by linear dynamics, and directed graphs model the communication network. First, new sufficient conditions for designing dynamic output feedback protocols of arbitrary order - including static output feedback as a particular case - are proposed for H∞ consensus for agents subject to external disturbances. The agents may experience parametric perturbations and communicate in uncertain polytopic topologies. Finally, we propose conditions that design protocols for time-varying output formation tracking. By considering a polytopic representation describing the network information, we present a procedure that allows the design of protocol gains for a family of topologies regardless of the number of agents. All proposed conditions are based on Linear Matrix Inequalities (LMIs). Numerical examples illustrate the effectiveness of the proposed approach.