自主车辆动态冲突控制可达集建模方法

Reachable Set Modeling Approach for Autonomous Vehicle's Dynamic Conflict Control

针对自主车辆主动安全控制建模问题,提出采用形式化建模与动态非合作博弈论相结合方法,研究多自主车辆动态冲突条件下轨迹追踪和碰撞躲避的控制策略。提出将系统控制划分为:轨迹追踪、碰撞躲避以及追踪与躲避切换的三种模式。考虑到自主车辆这个混成系统的复杂性与轨迹追踪偏差的存在性,提出对车辆碰撞躲避预警及控制采用可达集建模。为了保证车辆轨迹追踪的稳定性,提出具有迟滞特性的追踪躲避切换算法。系统验证表明,当阈值α_1=0.05和α_2=0.08时,提出的算法使系统稳定性最佳。而且系统控制的划分发挥了独立模型的优势,降低了计算成本。

Aiming at the problem of active safety control modeling for autonomous vehicles, a combined method using the formal modeling and the dynamic non-cooperative game theory was proposed for discussing a control strategy of trajectory tracking and collision avoidance under dynamic conflict condition between autonomous vehicles. And it was proposed to divide the system control into three modes trajectory tracki＇ng, collision avoidance and switching between trajectory tracking and collision avoidance. Taking into account the complexity as a hybrid systems and the existence of the trajectory tracking error for autonomous vehicle, a modeling approach based on reachable set was proposed for the collision avoidance control of vehicle. In addition, to ensure the stability of the trajectory tracking of vehicles, a switching algorithm with hysteresis characteristics was proposed. The system validation shows that the best stability is achieved for the proposed algorithm when the threshold al = 0.05 and a2 = 0.08. And the division of system control plays the advantage of independent model, which reduces the computational cost.