Li Auto（理想汽车）

Scaling Learning-based AEB with Massive Unlabeled Data

基于大规模无标签数据的可扩展学习型自动紧急制动

Xiangyu Wang, Yang Zhan, Mengxiang Hao, Chuanchuan Zhong, Yansong Jia, Junjie Zhang, Yu Han, Xin Jiang, Zhen Cao, Ying Wang, Yulun Song, Zhitao Xu

机构 * Li Auto 纠错

AI总结 提出稳定元反馈半监督学习框架，通过噪声感知解耦和运动学门控伪标签，利用大规模无标签数据提升自动紧急制动性能，实现超100:1正误触发比和35%无事故里程提升。

Comments Accepted for presentation at the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)