| In recent years,cooperative control of multi-agent systems has become a research hotspot in the fields of industrial control,artificial intelligence,and electronic information etc.Inmany practical situations,a single agent has been unable to solve complex,large-scale and uncertain problems.Using a single agent with high integration and structural complexity to perform corresponding tasks will inevitably result in high cost and low work efficiency.Therefore,cooperative control of multi-agent systems,a technology in which consists of multiple agents with relatively single functions combined and interacted according to certain rules to complete the global task,has a very broad application prospect due to its high efficiency and relatively low cost,such as drone formation performances,multi-robot cooperative assembly in production workshops or logistics sorting workshops,and distributed energy control in smart grids.This paper focuses on the consensus problem in the area of cooperative control of multi-agent systems,and studies the discrete-time multi-agent system subject to limited bandwidths of communication channels.First,for the first-order discrete-time multi-agent system with directed communication topologies,it considers the limited capacities of physical memories of agents,and proposes a distributed dynamic consensus protocol based on recursive channel filter.The distributed PI based consensus protocol is discussed as a special case,and the parameter conditions that can guarantee the multi-agent system achieve asymptotically consensus are obtained.Second,for the second-order discrete-time multi-agent system with undirected communication topologies,this paper proposes a consensus protocol by adopting channel non-recursive filters.The protocol is based on the data stored in the memory of the previous time,and the sensor is replaced by the memory to perform multi-state processing to save the amount of data quantization.The adopted channel non-recursive filter is designed by keeping the first two items of the general model of finite impulse non-recursive filter.In addition,this paper proposes an easy-to-implement simplified design method by using a single parameter,and discusses the consensus protocol design of non-recursive channel filters by using high-order implementations,and obtains the parameter conditions for the system to achieve asymptotic consensus.Finally,this paper describes the above closed-loop multi-agent system through a high-order state model,analyzes the influence of the probability quantizer,and proves that under all the proposed consensus protocols,the multi-agent system can achieve quantized consensus in the case of quantized communication,and obtain the parameter conditions that guarantee the multi-agent system achieve quantized consensus. |
