友好型人机协同转向鲁棒控制

Friendliness enhanced robust steering control based on driver-automation cooperative driving

中文关键词  智能汽车  人机共驾  T-S模糊控制  控制权博弈  驾驶人行为

英文关键词  intelligent vehicle  shared autonomy  T-S fuzzy control  game theory  driver behavior characteristics

中文摘要

      为了更好地让智辅系统“理解”驾驶人的转向控制意图,减少共驾过程中的人机冲突,本文提出了一种友好 型共享转向系统鲁棒控制算法.首先,建立了驾驶人在环的线性时变车路模型,对共驾型车道保持控制(LKAS)问题 进行了数学描述;其次针对复杂工况下道路曲率的不确定性、车辆动力学模型参数的时变特性以及由于车辆动力 学的强非线性存在而导致模型适配的不足,基于T-S模糊控制理论设计了状态反馈γ次优H∞鲁棒控制器;然后,设 计一种友好型人机控制权博弈模型,能够根据驾驶人工作负荷以及驾驶人状态来动态调整智辅系统参与驾驶的程 度, 实现了人机交互的显式表达和驾驶控制权的平稳动态分配;最后,基于驾驶人在环的Carsim/Simulink集成环境 对该共享控制算法的有效性与鲁棒性进行了验证,结果表明该共享控制算法能够较好地兼顾路径跟踪能力与驾驶 人的控制权裕度.

英文摘要

      In order to well-understand the driver’s steering control intention and reduce the conflict operations between the driver and intelligent system during the cooperative driving scenarios, a friendliness enhanced driver-automation shared steering control algorithm is thus proposed in this paper. Firstly, an augmented linear time-varying two degree-of-freedom vehicle model combined with driver-in-the-loop model is established for developing the shared lane-keeping assist system. Secondly, based on T-S fuzzy control theory, a state feedback γ suboptimal H∞ robust controller is designed to handle the uncertainty of road curvature caused by noise measurements, and also to cope with the time-varying characteristics of vehicle dynamics as well as model mismatch caused by strong nonlinearity of vehicle dynamics. Then, a novel game model for driver-automation control authority is designed. For the purpose of driver-automation friendliness, the degree of participation of the intelligent system during the process of cooperative driving is dynamically adjusted depending on the driver’s workload and the driving states of the driver, accordingly achieving a smoothly dynamic allocation of control authority between the driver and the intelligent system. Finally, the comparative experimental results successfully demonstrate that high tracking precision as well as the priority of the driver’s control authority can be well-balanced by the shared control proposed algorithm.