Automatic driving model training method and device combining GAIL and PPO

By combining GAIL and PPO algorithms, a multi-stage training framework and multi-task learning objective function are constructed, which solves the shortcomings of the autonomous driving model in driving behavior evaluation and strategy optimization, and achieves the improvement of the model's robustness and generalization capabilities, and is suitable for autonomous driving tasks in complex traffic environments.

CN120406192APending Publication Date: 2025-08-01ZHEJIANG WUWEN ZHIXING TECHNOLOGY CO LTD
0 Cites 4 Cited by

Patent Information

Application Number
CN202510370056.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-03-26
Publication Date
2025-08-01

AI Technical Summary

Technical Problem

The existing autonomous driving model training methods lack multi-scale driving behavior analysis and continuity evaluation, resulting in insufficient model robustness and generalization capabilities, and lack of effective multi-task learning mechanisms and adaptive confidence interval constraints for strategy optimization.

Method used

Combining generative adversarial imitation learning (GAIL) and proximity strategy optimization (PPO) algorithms, a multi-stage training framework is built, a dual-branch discriminator network is used for feature fusion, expert driving trajectory sequences are collected through a simulation environment, multi-task learning objective function is constructed, and adaptive confidence interval constraints and experience playback mechanisms are introduced.

Benefits of technology

It significantly improves the robustness and generalization capabilities of the autonomous driving model, can better handle complex scenarios, and generate safe and smooth driving strategies.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The embodiment of the invention provides an automatic driving model training method and device combining GAIL and PPO, and the method comprises the steps: collecting an expert driving track sequence through a simulation environment by designing a multi-stage training frame; a double-branch discriminator network is constructed, a frame-level branch analyzes single-frame behavior characteristics based on an attention mechanism, and a track-level branch adopts a recurrent neural network to process a continuous driving sequence so as to realize multi-scale characteristic fusion. Iterative training is carried out by adopting a near-end strategy optimization algorithm, a multi-task learning objective function is constructed, driving action generation and trajectory prediction are optimized at the same time, and an adaptive confidence interval constraint and experience playback mechanism is introduced. According to the method, the defects of the traditional technology in the aspects of driving behavior evaluation, strategy optimization and the like are effectively overcome, and the robustness and generalization ability of an automatic driving model are remarkably improved.
Need to check novelty before this filing date? Find Prior Art