TCP - Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline
Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline
Penghao Wu*, Xiaosong Jia*, Li Chen*, Junchi Yan, Hongyang Li, Yu Qiao
This repository contains the code for the paper Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline. We will release the code and dataset around Sept.
TCP is a simple unified framework to combine trajectory and control prediction for end-to-end autonomous driving. By time of release in June 17 2022, our method achieves new state-of-the-art on CARLA AD Leaderboard, in which we rank the first in terms of the Driving Score and Infraction Penalty using only a single camera as input.
If you find our repo or our paper useful, please use the following citation:
@article{wu2022trajectoryguided,
title={Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline},
author={Penghao Wu and Xiaosong Jia and Li Chen and Junchi Yan and Hongyang Li and Yu Qiao},
journal={arXiv preprint arXiv:2206.08129},
year={2022},
}
All code within this repository is under Apache License 2.0.