Distributed Observer-Based Consensus Over Directed Networks With Limited Communication Bandwidth Constraints

Abstract

In this paper, we present a distributed observer-type consensus controller design for general linear multiagent systems over directed digital networks with limited communication data rates. Considering that each agent can only exchange binary symbolic sequence with its neighbors due to bandwidth constraints, we adopt the probabilistic quantization approach in the distributed consensus controller design. Compared to the existing works, the main contribution is that we use distributed observer-type controllers to inhibit the impact of quantization noises in agent states. First, we study the observer-type protocols with total quantized communications, where the relative-state information of each agent is quantized. The quantized consensus can be achieved, and the upper bounds of steady-state errors in the mean-square sense are also characterized, which are related to the scale of networks, the control parameters, the quantization levels, and the structures of agents. Further, we extend the controller design based on total quantization to the partial quantization case where only the neighbors' information is quantized.