A knowledge-driven end-to-end autonomous driving method based on a sparse expert mechanism and a diffusion model

By using a sparse expert mechanism and a diffusion model, autonomous driving skills are divided into multiple expert modules, and an interpretable modular knowledge structure is constructed. This solves the problems of insufficient generalization ability and real-time performance of autonomous driving methods in changing environments, and enables efficient knowledge reuse and rapid task migration.

CN120716776BActive Publication Date: 2025-11-04TONGJI UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202511141736.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-08-15
Publication Date
2025-11-04
Estimated Expiration
2045-08-15

AI Technical Summary

Technical Problem

Existing autonomous driving methods suffer from insufficient generalization ability, poor transferability, and inadequate real-time performance when facing changing environments. They also lack clear knowledge modeling and scheduling mechanisms, resulting in difficulties in efficiently reusing existing knowledge and catastrophic forgetting problems when facing new tasks.

Method used

By employing a sparse expert mechanism and a diffusion model, driving skills are abstracted into multiple expert modules. These modules are then combined and scheduled using a sparse activation mechanism and a trajectory prediction diffusion model to construct an interpretable modular driving knowledge structure. Furthermore, incremental learning and expert reorganization and transfer for new tasks are achieved by freezing the original model and adding a new router.

Benefits of technology

It enables efficient knowledge organization and reuse in multiple scenarios, improves the generalization ability and real-time decision-making efficiency of autonomous driving systems, avoids catastrophic forgetting, adapts to dynamic environments, and supports rapid task migration.

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

Abstract

The present application relates to the field of intelligent automatic driving, and particularly relates to a knowledge-driven end-to-end automatic driving method based on a sparse expert mechanism and a diffusion model. The method comprises the following steps: S1: perception information processing and state coding; S2: sparse expert module construction and multi-task training; a sparse expert module composed of multiple experts is constructed, and reusable driving skills are obtained through multi-task behavior cloning training; S3: diffusion strategy network and action sequence generation; based on the diffusion model, a future multi-step control action sequence is generated from the current state condition, and a continuous and stable driving decision is formed; and S4: continuous learning and task migration mechanism. The present application constructs a modular driving knowledge structure that is combinable and interpretable, significantly improves the strategy modeling capability; the diffusion generation mechanism effectively improves the smoothness and stability of the decision-making process; the continuous learning and task migration mechanism with structural decoupling improves the long-term adaptability and deployment efficiency of the system.
Need to check novelty before this filing date? Find Prior Art