Multimodal Trajectory Prediction Conditioned on Lane-Graph Traversals

Nachiket Deo, Eric Wolff, Oscar Beijbom
Proceedings of the 5th Conference on Robot Learning, PMLR 164:203-212, 2022.

Abstract

Accurately predicting the future motion of surrounding vehicles requires reasoning about the inherent uncertainty in driving behavior. This uncertainty can be loosely decoupled into lateral (e.g., keeping lane, turning) and longitudinal (e.g., accelerating, braking). We present a novel method that combines learned discrete policy rollouts with a focused decoder on subsets of the lane graph. The policy rollouts explore different goals given current observations, ensuring that the model captures lateral variability. Longitudinal variability is captured by our latent variable model decoder that is conditioned on various subsets of the lane graph. Our model achieves state-of-the-art performance on the nuScenes motion prediction dataset, and qualitatively demonstrates excellent scene compliance. Detailed ablations highlight the importance of the policy rollouts and the decoder architecture.

Cite this Paper

BibTeX
@InProceedings{pmlr-v164-deo22a, title = {Multimodal Trajectory Prediction Conditioned on Lane-Graph Traversals}, author = {Deo, Nachiket and Wolff, Eric and Beijbom, Oscar}, booktitle = {Proceedings of the 5th Conference on Robot Learning}, pages = {203--212}, year = {2022}, editor = {Faust, Aleksandra and Hsu, David and Neumann, Gerhard}, volume = {164}, series = {Proceedings of Machine Learning Research}, month = {08--11 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v164/deo22a/deo22a.pdf}, url = {https://proceedings.mlr.press/v164/deo22a.html}, abstract = {Accurately predicting the future motion of surrounding vehicles requires reasoning about the inherent uncertainty in driving behavior. This uncertainty can be loosely decoupled into lateral (e.g., keeping lane, turning) and longitudinal (e.g., accelerating, braking). We present a novel method that combines learned discrete policy rollouts with a focused decoder on subsets of the lane graph. The policy rollouts explore different goals given current observations, ensuring that the model captures lateral variability. Longitudinal variability is captured by our latent variable model decoder that is conditioned on various subsets of the lane graph. Our model achieves state-of-the-art performance on the nuScenes motion prediction dataset, and qualitatively demonstrates excellent scene compliance. Detailed ablations highlight the importance of the policy rollouts and the decoder architecture.} }
Endnote
%0 Conference Paper %T Multimodal Trajectory Prediction Conditioned on Lane-Graph Traversals %A Nachiket Deo %A Eric Wolff %A Oscar Beijbom %B Proceedings of the 5th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2022 %E Aleksandra Faust %E David Hsu %E Gerhard Neumann %F pmlr-v164-deo22a %I PMLR %P 203--212 %U https://proceedings.mlr.press/v164/deo22a.html %V 164 %X Accurately predicting the future motion of surrounding vehicles requires reasoning about the inherent uncertainty in driving behavior. This uncertainty can be loosely decoupled into lateral (e.g., keeping lane, turning) and longitudinal (e.g., accelerating, braking). We present a novel method that combines learned discrete policy rollouts with a focused decoder on subsets of the lane graph. The policy rollouts explore different goals given current observations, ensuring that the model captures lateral variability. Longitudinal variability is captured by our latent variable model decoder that is conditioned on various subsets of the lane graph. Our model achieves state-of-the-art performance on the nuScenes motion prediction dataset, and qualitatively demonstrates excellent scene compliance. Detailed ablations highlight the importance of the policy rollouts and the decoder architecture.
APA
Deo, N., Wolff, E. & Beijbom, O.. (2022). Multimodal Trajectory Prediction Conditioned on Lane-Graph Traversals. Proceedings of the 5th Conference on Robot Learning, in Proceedings of Machine Learning Research 164:203-212 Available from https://proceedings.mlr.press/v164/deo22a.html.

Related Material