In this paper we propose a novel end-to-end learnable network that performs joint perception, prediction and motion planning for self-driving vehicles and produces interpretable intermediate representations. Unlike existing neural motion planners, our motion planning costs are consistent with our perception and prediction estimates. This is achieved by a novel differentiable semantic occupancy representation that is explicitly used as cost by the motion planning process. Our network is learned end-to-end from human demonstrations. The experiments in a large-scale manual-driving dataset and closed-loop simulation show that the proposed model significantly outperforms state-of-the-art planners in imitating the human behaviors while producing much safer trajectories.
@article{arxiv.2008.05930,
title = {Perceive, Predict, and Plan: Safe Motion Planning Through Interpretable Semantic Representations},
author = {Abbas Sadat and Sergio Casas and Mengye Ren and Xinyu Wu and Pranaab Dhawan and Raquel Urtasun},
journal= {arXiv preprint arXiv:2008.05930},
year = {2020}
}
Comments
European Conference on Computer Vision (ECCV) 2020