Synchronized model matching: a novel approach to cooperative control of nonlinear multi-agent systems

We present a novel hierarchical approach to cooperative control of multi-agent systems. The agents are modeled as non-identical nonlinear single-input single-output systems. The control strategy achieves synchronization of the agents to a common output trajectory of a desired type. It is based on synchronization of reference models on the network level and asymptotic model matching control on the agent level. In order to cooperatively attenuate disturbances acting on individual agents, we establish feedback from the agent level to the network level and introduce integral action on the network level. The approach is illustrated by a simulation example with four magnetic levitation systems.