Quantized consensus of multi-agent systems under switching topology

This paper investigates the upper bound on the acceptable time duration of broken communication links of a multi-agent system, whose consensus is expected. For the concerned multi-agent system, the state of each agent is quantized before transmissions. In order to save network resources, we propose a simplified control scheme named conservative input. When the state of an agent is higher(or lower) than the average of states of multiple agents of this system and it has negative(or positive) input, then its input will not be updated. Simulations confirm that the proposed conservative input well reduces the updating frequency of the control law while the system performance is only slightly degraded. Considering practical applications, we also analyze the multi-agent systems with bounded noise and show through simulations our conservative control can still maintain the desired consensus even in the presence of bounded noise.