Roads have well defined geometries, topologies, and traffic rules. While this
has been widely exploited in motion planning methods to produce maneuvers that
obey the law, little work has been devoted to utilize these priors in
perception and motion forecasting methods. In this paper we propose to
incorporate these structured priors as a loss function. In contrast to imposing
hard constraints, this approach allows the model to handle non-compliant
maneuvers when those happen in the real world. Safe motion planning is the end
goal, and thus a probabilistic characterization of the possible future
developments of the scene is key to choose the plan with the lowest expected
cost. Towards this goal, we design a framework that leverages REINFORCE to
incorporate non-differentiable priors over sample trajectories from a
probabilistic model, thus optimizing the whole distribution. We demonstrate the
effectiveness of our approach on real-world self-driving datasets containing
complex road topologies and multi-agent interactions. Our motion forecasts not
only exhibit better precision and map understanding, but most importantly
result in safer motion plans taken by our self-driving vehicle. We emphasize
that despite the importance of this evaluation, it has been often overlooked by
previous perception and motion forecasting works.
Description
[2006.02636] The Importance of Prior Knowledge in Precise Multimodal Prediction
%0 Journal Article
%1 casas2020importance
%A Casas, Sergio
%A Gulino, Cole
%A Suo, Simon
%A Urtasun, Raquel
%D 2020
%K bayesian priors
%T The Importance of Prior Knowledge in Precise Multimodal Prediction
%U http://arxiv.org/abs/2006.02636
%X Roads have well defined geometries, topologies, and traffic rules. While this
has been widely exploited in motion planning methods to produce maneuvers that
obey the law, little work has been devoted to utilize these priors in
perception and motion forecasting methods. In this paper we propose to
incorporate these structured priors as a loss function. In contrast to imposing
hard constraints, this approach allows the model to handle non-compliant
maneuvers when those happen in the real world. Safe motion planning is the end
goal, and thus a probabilistic characterization of the possible future
developments of the scene is key to choose the plan with the lowest expected
cost. Towards this goal, we design a framework that leverages REINFORCE to
incorporate non-differentiable priors over sample trajectories from a
probabilistic model, thus optimizing the whole distribution. We demonstrate the
effectiveness of our approach on real-world self-driving datasets containing
complex road topologies and multi-agent interactions. Our motion forecasts not
only exhibit better precision and map understanding, but most importantly
result in safer motion plans taken by our self-driving vehicle. We emphasize
that despite the importance of this evaluation, it has been often overlooked by
previous perception and motion forecasting works.
@article{casas2020importance,
abstract = {Roads have well defined geometries, topologies, and traffic rules. While this
has been widely exploited in motion planning methods to produce maneuvers that
obey the law, little work has been devoted to utilize these priors in
perception and motion forecasting methods. In this paper we propose to
incorporate these structured priors as a loss function. In contrast to imposing
hard constraints, this approach allows the model to handle non-compliant
maneuvers when those happen in the real world. Safe motion planning is the end
goal, and thus a probabilistic characterization of the possible future
developments of the scene is key to choose the plan with the lowest expected
cost. Towards this goal, we design a framework that leverages REINFORCE to
incorporate non-differentiable priors over sample trajectories from a
probabilistic model, thus optimizing the whole distribution. We demonstrate the
effectiveness of our approach on real-world self-driving datasets containing
complex road topologies and multi-agent interactions. Our motion forecasts not
only exhibit better precision and map understanding, but most importantly
result in safer motion plans taken by our self-driving vehicle. We emphasize
that despite the importance of this evaluation, it has been often overlooked by
previous perception and motion forecasting works.},
added-at = {2020-06-05T11:32:44.000+0200},
author = {Casas, Sergio and Gulino, Cole and Suo, Simon and Urtasun, Raquel},
biburl = {https://www.bibsonomy.org/bibtex/25c1852ad7a30de1e1607c1170064176f/kirk86},
description = {[2006.02636] The Importance of Prior Knowledge in Precise Multimodal Prediction},
interhash = {29842a5c0fdfaaad3a1197b2c0257f1d},
intrahash = {5c1852ad7a30de1e1607c1170064176f},
keywords = {bayesian priors},
note = {cite arxiv:2006.02636},
timestamp = {2020-06-05T11:32:44.000+0200},
title = {The Importance of Prior Knowledge in Precise Multimodal Prediction},
url = {http://arxiv.org/abs/2006.02636},
year = 2020
}