First-Order Dynamic Quantized Consensus of Multi-Agent Systems

Abstract

This paper analyses the consensus of Multi-Agent systems in the presence of imperfect information exchange between the agents. In particular, the agents exchange the quantized values of the relative state measurements instead of the relative values directly. Departing from the literature, the communication between any pair of agents can have a different quantization error. For a static communication graph, we derive the quantized consensus region. Assuming we can modify the communication graph topology, we propose a novel graph-switching approach to obtain a smaller quantized consensus region. The proposed switching graph topology controller utilizes fewer accurate edges while ensuring a better consensus region. Further, suppose the communication graph cannot be modified, assuming the quantization error of the edges can be modified. In that case, we propose a novel dynamic quantization scheme to achieve the best-quantized consensus region. We perform numerical simulations to validate the results presented.