Driver intention prediction seeks to anticipate drivers' actions by analyzing their behaviors with respect to surrounding traffic environments. Existing approaches primarily focus on late-fusion techniques, and neglect the importance of maintaining consistency between predictions and prevailing driving contexts. In this paper, we introduce a new framework called Cross-View Episodic Memory Transformer (CEMFormer), which employs spatio-temporal transformers to learn unified memory representations for an improved driver intention prediction. Specifically, we develop a spatial-temporal encoder to integrate information from both in-cabin and external camera views, along with episodic memory representations to continuously fuse historical data. Furthermore, we propose a novel context-consistency loss that incorporates driving context as an auxiliary supervision signal to improve prediction performance. Comprehensive experiments on the Brain4Cars dataset demonstrate that CEMFormer consistently outperforms existing state-of-the-art methods in driver intention prediction.
@article{arxiv.2305.07840,
title = {CEMFormer: Learning to Predict Driver Intentions from In-Cabin and External Cameras via Spatial-Temporal Transformers},
author = {Yunsheng Ma and Wenqian Ye and Xu Cao and Amr Abdelraouf and Kyungtae Han and Rohit Gupta and Ziran Wang},
journal= {arXiv preprint arXiv:2305.07840},
year = {2023}
}