Comprehensive preview decision-making method for direction and speed of intelligent vehicle based on rules and mechanisms

Abstract

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.