Research On The Formation Of Intelligent Vehicles Based On Network

In recent years, as the development of communication technology, sensortechnology and computer technology, multi-agent coordination control technology isbecoming an active research topic. With broad application prospects, Multi-agentcoordination control has been widely used in the military, transportation, aerospace andmarine exploration and other fields.Consistency problem is a basic problem of multi-agent coordination control, whichhas been studied for some years in history. Since the communication among agents inthe networked multi-agent systems is achieved by a network, it is inevitable that thenetwork-induced delay will occur while exchanging data among devices. However, timedelay can degrade the performance of control systems and even destabilize the system.The consensus problem for the multi-agent system with a constant communication timedelayis discussed in this paper. A new distributed consensus protocol is proposed byusing networked predictive control schemes. Necessary and sufficient conditions ofachieving consensus are obtained, which indicates that consensus is independent of thecommon constant delay. Compared to the other consensus protocol, an advantage of thispaper is that the proposed consensus protocol overcomes the effects of network delayactively rather than passively.Based on the theoretical research, this paper built an experimental platform formulti-agent coordination control and develop the VideoCapture visual positioningsystem, PC/104networked control systems. Visual positioning system is responsible forobtaining the coordinates of the intelligent car. The image is calibrated with animproved image calibration model and segmented with an adaptive thresholdsegmentation method. It’s indicated that the visual positioning system can meet therequirements of the experiment. PC/104networked control system is responsible forbuilding the Simulink block diagram in the host computer into an executable programand sending the real-time data of each parameter to smart car.Finally, an experiment of multi-agent coordination control was made in theexperimental platform which showed that the experimental platform can meet therequirements of theory study about the multi-agent system coordination control.