轨迹预测的不确定性会引发汽车换道避撞系统的预期功能安全(Safety of the intended functionality,SOTIF)风险,为此提出一种基于轨迹预测安全边界的智能汽车换道避撞控制方法.针对前方车辆紧急切入工况,首先利用切比雪夫区间分析方法量化前车切入轨迹的不确定性,推导轨迹预测安全边界的计算公式.然后据此进行自车换道碰撞危险分析,决策出换道避撞过程中自车的换道安全边界.接着以换道安全边界为约束,应用基于Tube的鲁棒模型预测控制(Tube-based robust model predictive control,Tube-RMPC)方法设计换道控制器,实现自车主动换道避撞控制.最后在Carsim与Simulink联合平台上开展仿真试验,比较此系统与没有考虑预测不确定性的换道避撞系统的安全性,并在硬件在环仿真台架上进行验证.结果表明,求解出的轨迹预测安全边界能有效包络前车切入轨迹的不确定性,基于轨迹预测安全边界的控制方法显著提升换道避撞系统的预期功能安全性能.
Intelligent Vehicle Lane Change Collision Avoidance System Based on the Safe Boundary of Trajectory Prediction
The uncertainty of trajectory prediction will lead to SOTIF risks of lane change collision avoidance(LCCA)system.To deal with it,a LCCA control method based on the safe boundary of trajectory prediction is proposed for intelligent vehicle.First,the uncertainty of the cut-in trajectory of the front vehicle is quantified by Chebyshey interval analysis method,and the formula of trajectory prediction safe boundary is constructed.On this basis,the collision risk analysis of lane change is carried out,and the vehicle lane change safe boundary in the process of lane change is determined.Then,with the lane change safe boundary as the constraint,a lane change controller based on Tube-RMPC is designed to realize active LCCA control of self-vehicle.The effectiveness of the proposed method is evaluated in Carsim/Simulink by comparing with ordinary LCCA system that does not account for prediction uncertainty.The safety of the system is further verified via the hardware-in-the-loop experiment.The results show that the obtained trajectory prediction safe boundary can effectively envelop the cut-in trajectories of the front vehicle and improve SOTIF of the LCCA system.
NETL
NSTL
万方数据
