Overview
The AutoRally platform is a high-performance testbed for self-driving vehicle research. The robot was developed at Georgia Tech by Brian Goldfain and Paul Drews, both advised by James Rehg, with contributions from many other students. A brief history of the project can be found here. The platform is designed as a self contained system that requires no external sensing or computing. The robot is a robust, cost-effective, and safe platform that opens the space of aggressive autonomous off-road driving to researchers and hobbyists.
The core software and simulation environment for the platform are publicly available along with tutorials. Build instructions for the chassis and compute box, complete parts list, CAD models, and operating procedures are released in a separate GitHub repository. If you are interested in the platform, fork and test the simulation, check out the documentation, and let us know if you have any questions.
Slides from Brian Goldfain's PhD thesis defense covering the platform and selected experiments is available in [PDF] and a 480MB [PPT] with videos.
Attribution
We request that any work that uses the AutoRally platform or its core components (e.g. simulation environments) please site the following article:
AutoRally: An Open Platform for Aggressive Autonomous Driving. Brian Goldfain, Paul Drews, Changxi You, Matthew Barulic, Orlin Velev, Panagiotis Tsiotras, James M. Rehg. Control Systems Magazine (CSM), 2019. [PDF][BibTex]
Publications that use AutoRally
Exploiting Singular Configurations for Controllable, Low-Power Friction Enhancement on Unmanned Ground Vehicles. Adam Foris, Nolan Wagener, Byron Boots, Anirban Mazumdar. IEEE Robotics and Automation Letters (RA-L), 2020. [PDF][BibTex]
Ensemble Bayesian Decision Making with Redundant Deep Perceptual Control Policies. Keuntaek Lee, Ziyi Wang, Bogdan Vlahov, Harleen Brar, Evangelos A Theodorou. To be appear, ICMLA, 2019. [PDF][BibTex]
Perceptual Attention-based Predictive Control. Keuntaek Lee, Gabriel Nakajima An, Viacheslav Zakharov, Evangelos A Theodorou. To be appear, CoRL, 2019. [PDF][BibTex]
Locally Weighted Regression Pseudo-Rehearsal for Online Learning of Vehicle Dynamics. Grady Williams, Brian Goldfain, Keuntaek Lee, Jason Gibson, James M Rehg, Evangelos A Theodorou. To be appear, CoRL, 2019. [PDF][BibTex]
An Online Learning Approach to Model Predictive Control. Nolan Wagener, Ching-An Cheng, Jacob Sacks, Byron Boots. Robotics: Science and Systems (RSS), 2019. [PDF][BibTex]
Early Failure Detection of Deep End-to-End Control Policy by Reinforcement Learning. Keuntaek Lee, Kamil Saigol, Evangelos A Theodorou. IEEE International Conference on Robotics and Automation (ICRA), 2019. [PDF][BibTex]
Vision-Based High-Speed Driving With a Deep Dynamic Observer. Paul Drews, Grady Williams, Brian Goldfain, Evangelos A Theodorou, James M Rehg. IEEE Robotics and Automation Letters (RA-L), 2019. [PDF][BibTex]
Robust Sampling Based Model Predictive Control with Sparse Objective Information. Grady Williams, Brian Goldfain, Paul Drews, Kamil Saigol, James M Rehg, Evangelos Theodorou. Robotics: Science and Systems (RSS), 2018. [PDF][BibTex]
Agile autonomous driving using end-to-end deep imitation learning. Yunpeng Pan, Ching-An Cheng, Kamil Saigol, Keuntaek Lee, Xinyan Yan, Evangelos Theodorou, Byron Boots. Robotics: Science and Systems (RSS), 2018. [PDF][BibTex]
Best Response Model Predictive Control for Agile Interactions Between Autonomous Ground Vehicles. Grady Williams, Brian Goldfain, Paul Drews, James M Rehg, Evangelos A Theodorou. IEEE International Conference on Robotics and Automation (ICRA), 2018. [PDF][BibTex]
Information-Theoretic Model Predictive Control: Theory and Applications to Autonomous Driving. Grady Williams, Paul Drews, Brian Goldfain, James M Rehg, Evangelos A Theodorou. IEEE Transactions on Robotics (T-RO), 2018. [PDF][BibTex]
Aggressive Deep Driving: Model Predictive Control with a CNN Cost Model. Paul Drews, Grady Williams, Brian Goldfain, Evangelos A. Theodorou, James M. Rehg. Conference on Robotic Learning (CoRL), 2017. [PDF][BibTex]
Information theoretic MPC for model-based reinforcement learning. Grady Williams, Paul Drews, Brian Goldfain, James M Rehg, Evangelos A Theodorou. IEEE International Conference on Robotics and Automation (ICRA), 2017. [PDF][BibTex]
Vehicle Modeling and Parameter Estimation Using Adaptive Limited Memory Joint-State UKF. Changxi You, Panagiotis Tsiotras. American Control Conference (ACC), 2017. [PDF][BibTex]
Aggressive driving with model predictive path integral control. Grady Williams, Paul Drews, Brian Goldfain, James M Rehg, Evangelos A Theodorou. IEEE International Conference on Robotics and Automation (ICRA), 2016. [PDF][BibTex]
Recent Posts
RSS 2018
RSS 2018 Papers and Videos
ICRA 2018
ICRA 2018 Paper and Video
ICRA 2017 Video and Paper
Our new video for information theoretic MPPI is avaialble
Build Instructions v1.3
Improvements and minor updates to AutoRally Platform documentation
Build Instructions v1.1
Complete build instructions and all supporting files are publicly available!
MPPI Explanation Video
Experimental results and explanation of the MPPI algorithm on the AutoRally platform
ICRA 2016 Trailer
ICRA 2016 publication results were included in a highlights video.