We are currently finalizing this repository to coincide with the upcoming IROS paper.
This repository contains the implementation of the Long-Short-Term Online Learning (LSTOL) method proposed in our IROS 2024 paper entitled Preventing Catastrophic Forgetting in Continuous Online Learning for Autonomous Driving.
Autonomous vehicles require online learning capabilities to enable long-term, unattended operation. However, long-term online learning is accompanied by the problem of forgetting previously learned knowledge. This paper introduces an online learning framework that includes a catastrophic forgetting prevention mechanism, named Long-Short-Term Online Learning (LSTOL). The framework consists of a set of short-term learners and a long-term controller, where the former is based on the concept of ensemble learning and aims to achieve rapid learning iterations, while the latter contains a simple yet efficient probabilistic decision-making mechanism combined with four control primitives to achieve effective knowledge maintenance. A novel feature of the proposed LSTOL is that it avoids forgetting while learning autonomously. In addition, LSTOL makes no assumptions about the model type of short-term learners and the continuity of the data. The effectiveness of the proposed framework is demonstrated through experiments across well-known datasets in autonomous driving, including KITTI and Waymo.
What needs to be installed?
How to run this stack?
If you find our code or paper useful for your research, please cite:
@inproceedings{yr24iros,
title={Preventing Catastrophic Forgetting in Continuous Online Learning for Autonomous Driving},
author={Rui Yang and Zhi Yan and Tao Yang and Tomas Krajnik and Yassine Ruichek},
booktitle={2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
pages={},
year={2024}
}