Discrete Residual Flow for Probabilistic Pedestrian Behavior Prediction

Ajay Jain, Sergio Casas, Renjie Liao, Yuwen Xiong, Song Feng, Sean Segal, Raquel Urtasun
Proceedings of the Conference on Robot Learning, PMLR 100:407-419, 2020.

Abstract

Self-driving vehicles plan around both static and dynamic objects, applying predictive models of behavior to estimate future locations of the objects in the environment. However, future behavior is inherently uncertain, and models of motion that produce deterministic outputs are limited to short timescales. Particularly difficult is the prediction of human behavior. In this work, we propose the discrete residual flow network (DRF-NET), a convolutional neural network for human motion prediction that captures the uncertainty inherent in long-range motion forecasting. In particular, our learned network effectively captures multimodal posteriors over future human motion by predicting and updating a discretized distribution over spatial locations. We compare our model against several strong competitors and show that our model outperforms all baselines.

Cite this Paper

BibTeX
@InProceedings{pmlr-v100-jain20a, title = {Discrete Residual Flow for Probabilistic Pedestrian Behavior Prediction}, author = {Jain, Ajay and Casas, Sergio and Liao, Renjie and Xiong, Yuwen and Feng, Song and Segal, Sean and Urtasun, Raquel}, booktitle = {Proceedings of the Conference on Robot Learning}, pages = {407--419}, year = {2020}, editor = {Kaelbling, Leslie Pack and Kragic, Danica and Sugiura, Komei}, volume = {100}, series = {Proceedings of Machine Learning Research}, month = {30 Oct--01 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v100/jain20a/jain20a.pdf}, url = {https://proceedings.mlr.press/v100/jain20a.html}, abstract = {Self-driving vehicles plan around both static and dynamic objects, applying predictive models of behavior to estimate future locations of the objects in the environment. However, future behavior is inherently uncertain, and models of motion that produce deterministic outputs are limited to short timescales. Particularly difficult is the prediction of human behavior. In this work, we propose the discrete residual flow network (DRF-NET), a convolutional neural network for human motion prediction that captures the uncertainty inherent in long-range motion forecasting. In particular, our learned network effectively captures multimodal posteriors over future human motion by predicting and updating a discretized distribution over spatial locations. We compare our model against several strong competitors and show that our model outperforms all baselines.} }
Endnote
%0 Conference Paper %T Discrete Residual Flow for Probabilistic Pedestrian Behavior Prediction %A Ajay Jain %A Sergio Casas %A Renjie Liao %A Yuwen Xiong %A Song Feng %A Sean Segal %A Raquel Urtasun %B Proceedings of the Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2020 %E Leslie Pack Kaelbling %E Danica Kragic %E Komei Sugiura %F pmlr-v100-jain20a %I PMLR %P 407--419 %U https://proceedings.mlr.press/v100/jain20a.html %V 100 %X Self-driving vehicles plan around both static and dynamic objects, applying predictive models of behavior to estimate future locations of the objects in the environment. However, future behavior is inherently uncertain, and models of motion that produce deterministic outputs are limited to short timescales. Particularly difficult is the prediction of human behavior. In this work, we propose the discrete residual flow network (DRF-NET), a convolutional neural network for human motion prediction that captures the uncertainty inherent in long-range motion forecasting. In particular, our learned network effectively captures multimodal posteriors over future human motion by predicting and updating a discretized distribution over spatial locations. We compare our model against several strong competitors and show that our model outperforms all baselines.
APA
Jain, A., Casas, S., Liao, R., Xiong, Y., Feng, S., Segal, S. & Urtasun, R.. (2020). Discrete Residual Flow for Probabilistic Pedestrian Behavior Prediction. Proceedings of the Conference on Robot Learning, in Proceedings of Machine Learning Research 100:407-419 Available from https://proceedings.mlr.press/v100/jain20a.html.

Related Material