Research On Consensus Problems Of Multi-agent System Based On Sampled-data Control And Quantized Communication

Due to the great advantage of distributed Control and the rapid development of Network Technology, Network-Based distributed coordination control become a research hot in control field. The consensus problem is the basis for distributed coordination control of multi-agent system. The consensus of Multi-agent system means that every agent interact information through communication network to adjust their behavior. Eventually the whole system tend to the asymptotic consensus in one or more state quantity.In many practical cases, despite the control quantity is a continuous process, but the controller and the network often can’t complete computing and sending message in arbitrary high speed. The controller can only synthetically process the information of the sampling times, the system can only get the information of the discrete sampling point. The applied control is discrete time sampling point, when the sampled-data interact between agents through the network, because of the limitation of network bandwidth and communication rate, the information exchange between Multi-agent System is not entirely accurate. So the Sampled-Data need to be quantified before transferring. Therefore, in this paper, the Consensus problems of Multi-agent System Based on Sampled-Data Control and Quantized communication is studied. Detailed content of this paper is as follows:1. Assuming that the information exchange between Multi-agent System is entirely accurate, the sampled-data consensus problem of second-order multi-agent under the condition of fixed topology is studied. Theoretical analys is achieve the sufficient and necessary conditions and give the sampling period and control parameters in order to realize sampled-data consensus. MATLAB simulation prove the validity of the algorithm, so that the multi-agent consensus algorithm have better practical value.2. In practical cases, because of the limitation of network bandwidth and communication rate, the information exchange between Multi-agent System is not entirely accurate. Considering the effect of sampling and quantization on consensus problem at the same time, the quantized consensus problem of second-order multi-agent based on sampled-data is studied, designing protocol for quantized consensus. Theoretical analysis achieve the sufficient and necessary conditions of and give the sampling period and control parameters in order to realize quantized consensus based on sampled-data. The condition ensure the final consensus error converge to a small neighborhood of the origin. When the quantization interval tends to zero, the consensus error will converge to zero. MATLAB simulation prove the validity of the algorithm.3.At present, the research on multi-agent remain in theoretical research stage. And the study of experimental platform for consensus of multi-agent is less. Using the above theory, this paper established consensus speed control experiment platform of multi-agent car system based on the zigbee network. Every car interact information through zigbee communication network. After consensus coupling operation, these car achieve speed consistency, which proves the consensus theory and realize the practical application.