Controller Design and Experimental Validation for Connected Vehicle Systems Subject to Digital Effects and Stochastic Packet Drops

connected automated vehicles;robots;stochastic delays;string stability;V2X communication;plant stability;Mechanical Engineering;Engineering;Mechanical Engineering

Qin, WubingOldham, Kenn Richard ;

University of Michigan

关键词: connected automated vehicles;    robots;    stochastic delays;    string stability;    V2X communication;    plant stability;    Mechanical Engineering;    Engineering;    Mechanical Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/145874/wubing_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Vehicle-to-everything (V2X) communication allows vehicles to monitor the nearby traffic environment, including participants that are beyond the line of sight. Equipping conventional vehicles with V2X devices results in connected vehicles (CVs) while incorporating the information provided by V2X devices into the controllers of automated vehicles (AVs) leads to connected automated vehicles (CAVs). CAVs have great potential for improving driving comfort, reducing fuel consumption and advancing active safety for individual vehicles, as well as enhancing traffic efficiency and mobility for human-dominated traffic systems. In this dissertation, we study a class of connected cruise control (CCC) algorithms for longitudinal control of CAVs, where they respond to the motion information of one or multiple connected vehicles ahead. For validation and demonstration purposes, we utilize a scaled connected vehicle testbed consisting of a group of ground robots, which can provide us with insights about the controller design of full-size vehicles.On the one hand, intermittencies in V2X communication combined with the digital implementation of controllers introduce information delays. To ensure the performance of individual CAVs and the overall traffic, a set of methods is proposed for design and analysis of such communication-based controllers. We validate them with the scaled testbed by conducting a series of experiments on two-car predecessor-follower systems, cascaded predecessor-follower systems, and more complex connected vehicle systems. It is demonstrated that CAVs utilizing information about multiple preceding vehicles in the CCC algorithm can improve the system performance even for low penetration levels. This can be beneficial at the early stage of vehicle automation when human-driven vehicles still dominate the traffic system.On the other hand, we study the delay variations caused by stochastic packet drops in V2X communication and derive the stochastic processes describing the dynamics for the predecessor-follower systems. The dynamics of the mean, second moment and covariance are utilized to obtain stability conditions. Then the results of the two-car predecessor-follower system with stochastic delay variations are extended to an open chain as well as to a closed ring of cascaded predecessor-followers where stochastic packet drops lead to heterogeneity among different V2X devices. It is shown that the proposed analytical methods allow CCC design for CAVs that can achieve stability and stochastic disturbance attenuation in the presence of stochastic packet drops in complex connected vehicle systems.