RAP: Risk-Aware Prediction for Robust Planning

Haruki Nishimura, Jean Mercat, Blake Wulfe, Rowan Thomas McAllister, Adrien Gaidon
Proceedings of The 6th Conference on Robot Learning, PMLR 205:381-392, 2023.

Abstract

Robust planning in interactive scenarios requires predicting the uncertain future to make risk-aware decisions. Unfortunately, due to long-tail safety-critical events, the risk is often under-estimated by finite-sampling approximations of probabilistic motion forecasts. This can lead to overconfident and unsafe robot behavior, even with robust planners. Instead of assuming full prediction coverage that robust planners require, we propose to make prediction itself risk-aware. We introduce a new prediction objective to learn a risk-biased distribution over trajectories, so that risk evaluation simplifies to an expected cost estimation under this biased distribution. This reduces sample complexity of the risk estimation during online planning, which is needed for safe real-time performance. Evaluation results in a didactic simulation environment and on a real-world dataset demonstrate the effectiveness of our approach. The code and a demo are available.

Cite this Paper

BibTeX
@InProceedings{pmlr-v205-nishimura23a, title = {RAP: Risk-Aware Prediction for Robust Planning}, author = {Nishimura, Haruki and Mercat, Jean and Wulfe, Blake and McAllister, Rowan Thomas and Gaidon, Adrien}, booktitle = {Proceedings of The 6th Conference on Robot Learning}, pages = {381--392}, year = {2023}, editor = {Liu, Karen and Kulic, Dana and Ichnowski, Jeff}, volume = {205}, series = {Proceedings of Machine Learning Research}, month = {14--18 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v205/nishimura23a/nishimura23a.pdf}, url = {https://proceedings.mlr.press/v205/nishimura23a.html}, abstract = {Robust planning in interactive scenarios requires predicting the uncertain future to make risk-aware decisions. Unfortunately, due to long-tail safety-critical events, the risk is often under-estimated by finite-sampling approximations of probabilistic motion forecasts. This can lead to overconfident and unsafe robot behavior, even with robust planners. Instead of assuming full prediction coverage that robust planners require, we propose to make prediction itself risk-aware. We introduce a new prediction objective to learn a risk-biased distribution over trajectories, so that risk evaluation simplifies to an expected cost estimation under this biased distribution. This reduces sample complexity of the risk estimation during online planning, which is needed for safe real-time performance. Evaluation results in a didactic simulation environment and on a real-world dataset demonstrate the effectiveness of our approach. The code and a demo are available.} }
Endnote
%0 Conference Paper %T RAP: Risk-Aware Prediction for Robust Planning %A Haruki Nishimura %A Jean Mercat %A Blake Wulfe %A Rowan Thomas McAllister %A Adrien Gaidon %B Proceedings of The 6th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2023 %E Karen Liu %E Dana Kulic %E Jeff Ichnowski %F pmlr-v205-nishimura23a %I PMLR %P 381--392 %U https://proceedings.mlr.press/v205/nishimura23a.html %V 205 %X Robust planning in interactive scenarios requires predicting the uncertain future to make risk-aware decisions. Unfortunately, due to long-tail safety-critical events, the risk is often under-estimated by finite-sampling approximations of probabilistic motion forecasts. This can lead to overconfident and unsafe robot behavior, even with robust planners. Instead of assuming full prediction coverage that robust planners require, we propose to make prediction itself risk-aware. We introduce a new prediction objective to learn a risk-biased distribution over trajectories, so that risk evaluation simplifies to an expected cost estimation under this biased distribution. This reduces sample complexity of the risk estimation during online planning, which is needed for safe real-time performance. Evaluation results in a didactic simulation environment and on a real-world dataset demonstrate the effectiveness of our approach. The code and a demo are available.
APA
Nishimura, H., Mercat, J., Wulfe, B., McAllister, R.T. & Gaidon, A.. (2023). RAP: Risk-Aware Prediction for Robust Planning. Proceedings of The 6th Conference on Robot Learning, in Proceedings of Machine Learning Research 205:381-392 Available from https://proceedings.mlr.press/v205/nishimura23a.html.

Related Material