A split control method for team cooperative driving

Abstract

The invention relates to a control method for fleet cooperative driving splitting. A fleet splitting behavior rule library is established in advance in the controller of an automatic driving vehicle; If the distance between the self-driving vehicle and the front vehicle in the cruising convoy meets the separation distance, the team split event is triggered, and the controller of the self-driving vehicle controls the self-driving vehicle to perform the team split driving behavior; when performing the team split driving behavior, The controller of the self-driving vehicle obtains the pose information of the self-driving vehicle, and predicts the movement action of the self-driving vehicle at the next moment; then judges whether the movement action of the self-driving vehicle at the next moment matches the fleet split behavior rule base, If it matches, use the pose information of the current self-driving vehicle to iteratively update the fleet split driving behavior rules in the fleet split behavior rule base. The invention can realize optimal control effect on fleet splitting.

Description

technical field [0001] The invention belongs to the technical field of vehicle automatic driving, and in particular relates to a split control method for team cooperative driving. Background technique [0002] Self-driving vehicles are not only important electromechanical products, but also carriers of new energy, new materials, new equipment and other high-tech technologies, covering basic science and common technical issues in cross-cutting fields such as environmental perception, planning and decision-making, information communication, and automatic control. Among them, planning and decision-making are responsible for generating safe and reasonable driving behaviors like skilled drivers, so as to control the vehicle automatically. [0003] With the development of artificial intelligence technology represented by deep learning and machine learning, the "end-to-end" driving decision-making method that simulates the driver to directly generate driving decision-making instruc...

AI Technical Summary

Problems solved by technology

However, the offline training of this method in the real scene requires a large enough sample, and the provided samples often contain concerns that are irrelevant to driving decisions, and the interpretability is poor; at the same time, the test in the simulation scene cannot be directly used in the actual environment, lacking Practicality

Images