On The Path Tracking And Cooperation Control Of The Multiple Autonomous Vehicle System

The research results of multiple autonomous vehicles system can effectively improve the existing traffic conditions, ease the traffic pressure, improve the road safety and promote the development of intelligent driving. In recent years, it has been one of the hot researches.In this paper, we build an indoor multi autonomous vehicle platform for related theoretical studies, including the path tracking experiment on the platform, the safety control in no signal light intersections and task allocation in the urban traffic network. The main studies are as follows:(1) A multiple autonomous vehicle collaboration platform is build in this paper. We design the architecture of the entire platform, debug the indoor positioning system, include the coverage and positioning accuracy; build wireless communications networks and standardized the way of information transmitting. We also design the autonomous vehicles subsystem and debug the related parts, including infrared sensors and the camera. On this basis, the software platform is developed, so that we can network the platform modules, rich the functions of the platform and realize the information real-time interaction inside the platform.(2) An electronic map library is established for autonomous vehicles. The desired trajectory is stored in the form of target points. The autonomous vehicle combines indoor position information and real-time kinematic model to estimation the state. A controller is designed to complete the path tracking and it is tested on the platform we build?(3) Secure collaboration for collision avoidance in typical intersection is studied in this work, we use the hybrid automata model for multiple autonomous vehicle system modeling. According to the status information of vehicles, we determine the dangerous areas in the displacement coordinate plane of two autonomous vehicles. And in the cases that the two vehicles'initial positions are outside or inside the danger zone, we design the different collision avoidance control strategies. Simulation results also demonstrate the effectiveness of the algorithm.(4) Under the background of goods transport, we simulate the real road environment and use the second-price sealed-bid auction for task allocation. Further, we introduce the ability and preference of autonomous vehicles to reduce bidding conflict. To the newly added tasks, we proposed redistribution to optimize existing allocation scheme and further optimize the overall performance of the system. Simulation experiments for different number of vehicles and tasks are carried.In summary, the main work of this paper is setting up the whole real-time simulation platform to provide a good platform support, including both hardware and software parts and make simulation and experimental analysis on path tracking, safety collaboration and dynamic task allocation.