On Exposing the Challenging Long Tail in Future Prediction of Traffic Actors
This work addresses a critical safety issue for autonomous systems by enhancing prediction in rare, complex traffic scenarios, representing an incremental improvement over existing methods.
The paper tackles the problem of predicting future states of traffic actors, focusing on improving performance on infrequent but critical scenarios that are often suboptimally handled by common losses. The proposed method supplements the usual loss with one that clusters challenging cases in embedding space, leading to improved performance on these scenarios across four public datasets while maintaining overall stability.
Predicting the states of dynamic traffic actors into the future is important for autonomous systems to operate safelyand efficiently. Remarkably, the most critical scenarios aremuch less frequent and more complex than the uncriticalones. Therefore, uncritical cases dominate the prediction. In this paper, we address specifically the challenging scenarios at the long tail of the dataset distribution. Our analysis shows that the common losses tend to place challenging cases suboptimally in the embedding space. As a consequence, we propose to supplement the usual loss with aloss that places challenging cases closer to each other. This triggers sharing information among challenging cases andlearning specific predictive features. We show on four public datasets that this leads to improved performance on the challenging scenarios while the overall performance stays stable. The approach is agnostic w.r.t. the used network architecture, input modality or viewpoint, and can be integrated into existing solutions easily. Code is available at https://github.com/lmb-freiburg/Contrastive-Future-Trajectory-Prediction