复杂环境下车辆自主决策可达集分析

An analysis of vehicle autonomous decision-making reachable set in complex environment

针对传统的车辆安全性验证模型多为单目标简单环境而导致其适应性低的问题,提出基于哈密顿-雅可比方程的可达集分析方法.在经典超车问题的建模中,通过加入对向来车,增加模型的复杂度和可达集计算的约束条件,识别并预测场景中各交通参与者的机动行为.为了提高计算的准确性,采用水平集方法描述可达集演变过程,引入曲线行驶下车辆行驶特性函数fa(x),以适应自主车辆在不同行为决策下可达集的变化.仿真结果表明,该方法在复杂的道路交通环境下不仅能够判断自主车辆目前状态的安全性,而且能够预测未来时刻的安全性,同时可为智能驾驶决策的安全性提供参考依据.

Aiming at the problem that the vehicle safety verification model is less adaptable than the single target and so on,this paper puts forward a Jacobian analysis method based on Hamiltonian equation,and adds the complexity of the model and the constraint conditions of the calculation of the models in the modeling of the classical overtaking problem to identify and predict the maneuvering behavior of each traffic participant in the scene.In order to improve the accuracy of the calculation,the level set method is used to describe the evolution process of the set,and the driving characteristic function fa(x)of the curve is introduced to adapt the different behavior decision-making of autonomous vehicles.The results show that this method can not only judge the safety of autonomous vehicles in the present condition,but also predict the security of future time,and provide reference for the safety of intelligent driving decision.