According to the existing driver model, the objective function is coupled with continuity factors and discontinuity factors, which makes it difficult to determine the weighting coefficients in multi-objective optimization, which will cause dangerous situations such as the vehicle rushing out of the road boundary; in response to this problem, this paper proposes a driver model adapted to the complex traffic environment, based on the mechanism modeling of continuity factors and rule modeling of discontinuity factors. In view of the difficulty of traditional optimization algorithms to find a balance between efficiency and accuracy, this paper proposes a grid optimization algorithm that takes into account both efficiency and accuracy. In order to reduce the amount of calculation in the preview decision-making process, this paper proposes a curve integral method based on the laws of vehicle kinematics to predict the position of the vehicle to judge whether a collision will occur. The driver model is established in the Simulink simulation environment, and the C-level prototype model in the vehicle dynamics simulation software CarSim is selected as the control object, the results show that the proposed the preview decision model effectively solves the problem of divergence in the optimization solution, and can also ensure safety and traffic rules in a complex traffic environment, improving the quality of the model.

根据现有的驾驶员模型，目标函数耦合了连续性因素和不连续性因素，使得多目标优化中的权重系数难以确定，会造成车辆冲出道路边界等危险情况；针对这一问题，本文基于连续性因素的机理建模和不连续性因素的规则建模，提出了一种适应复杂交通环境的驾驶员模型。针对传统优化算法难以在效率和精度之间找到平衡点的问题，本文提出一种兼顾效率和精度的网格优化算法。为了减少预览决策过程中的计算量，本文提出了一种基于车辆运动学规律的曲线积分法，通过预测车辆的位置来判断是否会发生碰撞。在 Simulink 仿真环境中建立驾驶员模型，选取车辆动力学仿真软件 CarSim 中的 C 级原型模型作为控制对象，结果表明所提出的预览决策模型有效地解决了车辆动力学仿真中的发散问题。优化方案，在复杂的交通环境中也能保证安全和交通规则，提高模型质量。