Problem
Research questions and friction points this paper is trying to address.
learning-based control
vehicle dynamics
safety guarantees
curriculum learning
predictive safety filter
Predictive safety filter enhanced curriculum learning control for efficient vehicle dynamics controller
🤖 AI Summary
This work addresses the limitations of traditional vehicle dynamics control, which relies on labor-intensive parameter tuning, and purely learning-based approaches, which often lack safety guarantees and robustness. To bridge this gap, the paper proposes a novel framework that integrates curriculum learning with a physics-informed predictive safety filter, embedding predictive safety mechanisms directly into the curriculum learning pipeline for the first time. This integration enhances both stability and agility of the controller under complex driving conditions while rigorously ensuring safety. Implemented and validated in a closed-loop Python-CarSim environment, the proposed method demonstrates superior performance, scalability, and safety assurances across diverse driving scenarios compared to existing approaches.
📝 Abstract
Recent advances in learning-based control have enabled impressive achievements in solving complex control problems in various domains. However, since learning-based control may not be able to realize safety-guaranties, it is of great importance to enhance safety and robustness while maintaining good performances. Take vehicle motion \& dynamics control as an example, in order to overcome the pain points of traditional methods such as heavy parameter calibration effort and learning-based control to bring better performance and efficiency in stability \& agility over prior work for state-based vehicle control tasks, in this work, our method aims to develop a curriculum learning controller enhanced with physics-based predictive safety filter. The validation is conducted with the Python-CarSim platform, demonstrating better improvements and scalability under various maneuvers.
Innovation
Methods, ideas, or system contributions that make the work stand out.
predictive safety filter
curriculum learning
learning-based control
vehicle dynamics control
physics-based safety
🔎 Similar Papers
No similar papers found.
