Intelligent car following decision-making and control method based on cognitive risk balance

Abstract

The present application relates to a smart car following decision and control method based on cognitive risk balance. In the interval, if no, the function between longitudinal acceleration and cognitive risk is used to predict and obtain the response acceleration of human drivers under the action of the corresponding risk intensity, as the anthropomorphic response acceleration of the intelligent vehicle to deal with the corresponding risk, and the anthropomorphic response Acceleration is used to adjust the cognitive risk to keep it balanced; based on the response acceleration, the throttle opening and braking pressure intensity are predicted and obtained through the vehicle longitudinal inverse dynamics model, so as to adjust the vehicle speed in real time and realize the intelligent car following. As a result, the application of the anthropomorphic cognitive risk balance mechanism in the decision-making of the longitudinal driving behavior of smart cars is realized. By learning the driving decision-making rules of excellent drivers, it can guide the anthropomorphic following driving of smart cars.

Description

technical field [0001] This application relates to the technical field of smart car applications, in particular to a smart car follow-up decision-making and control method based on cognitive risk balance. Background technique [0002] For smart cars, such as in smart driving and automatic driving, the car-following behavior models used in its longitudinal dynamics motion decision-making mainly include: (1) Stimulus-response models, such as GM model (general motor) and OVM (Optimal velocity model) model ; (2) Safety distance model, such as Gipps model; (3) Social force model, such as IDM (Intelligent driver model) model and LCM (longitudinal control model) model; (4) Statistical learning model, such as neural network model, deep learning model and reinforcement learning models. [0003] The existing car-following models are usually empirical models or semi-empirical models, which have insufficient physical meaning, insufficient interpretability, poor transferability, and fai...

AI Technical Summary

Problems solved by technology

[0004] This application provides a smart car following decision-making and control method based on cognitive risk balance to solve the problem that the existing following models are usually empirical models or semi-empirical models, and the physical meaning is not clear enough, the explainability is not enough, and it can be transferred The lack of anthropomorphism of the car-following model due to the failure to fully explore the human driver's following behavior, and the failure to combine with the vehicle inverse dynamics model to achieve real-time and reliable throttle and brake pressure control. The application of the anthropomorphic cognitive risk balance mechanism in the longitudinal driving behavior decision-making of smart cars, by learning the driving decision-making rules of excellent drivers, to guide the anthropomorphic follow-up of smart cars

Images