Raising context awareness in motion forecasting

Hédi Ben-Younes, Éloi Zablocki, Mickaël Chen, Patrick Pérez, Matthieu Cord

CVPR Workshop on Autonomous Driving (WAD) 2022

Abstract

Learning-based trajectory prediction models have encountered great success, with the promise of leveraging contextual information in addition to motion history. Yet, we find that state-of-the-art forecasting methods tend to overly rely on the agent's current dynamics, failing to exploit the semantic contextual cues provided at its input. To alleviate this issue, we introduce CAB, a motion forecasting model equipped with a training procedure designed to promote the use of semantic contextual information. We also introduce two novel metrics, dispersion and convergence-to-range, to measure the temporal consistency of successive forecasts, which we found missing in standard metrics. Our method is evaluated on the widely adopted nuScenes Prediction benchmark as well as on a subset of the most difficult examples from this benchmark.

BibTeX

@inproceedings{cab2022,
  author    = {Hedi Ben{-}Younes and
               {\'{E}}loi Zablocki and
               Micka{\"{e}}l Chen and
               Patrick P{\'{e}}rez and
               Matthieu Cord},
  title     = {Raising context awareness in motion forecasting},
  booktitle = {CVPR Workshop on Autonomous Driving (WAD)},
  year      = {2022}
}