Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation
In this work, we present a data-driven simulation and training engine capable of learning end-to-end autonomous vehicle control policies using only sparse rewards. By leveraging real, human-collected trajectories through an environment, we render novel training data that allows virtual agents to dri...
| Main Authors: | , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2021
|
| Online Access: | https://hdl.handle.net/1721.1/130456 |
| _version_ | 1811083935024676864 |
|---|---|
| author | Amini, Alexander A Gilitschenski, Igor Phillips, Jacob Moseyko, Julia Banerjee, Rohan Karaman, Sertac Rus, Daniela L |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Amini, Alexander A Gilitschenski, Igor Phillips, Jacob Moseyko, Julia Banerjee, Rohan Karaman, Sertac Rus, Daniela L |
| author_sort | Amini, Alexander A |
| collection | MIT |
| description | In this work, we present a data-driven simulation and training engine capable of learning end-to-end autonomous vehicle control policies using only sparse rewards. By leveraging real, human-collected trajectories through an environment, we render novel training data that allows virtual agents to drive along a continuum of new local trajectories consistent with the road appearance and semantics, each with a different view of the scene. We demonstrate the ability of policies learned within our simulator to generalize to and navigate in previously unseen real-world roads, without access to any human control labels during training. Our results validate the learned policy onboard a full-scale autonomous vehicle, including in previously un-encountered scenarios, such as new roads and novel, complex, near-crash situations. Our methods are scalable, leverage reinforcement learning, and apply broadly to situations requiring effective perception and robust operation in the physical world. |
| first_indexed | 2024-09-23T12:41:57Z |
| format | Article |
| id | mit-1721.1/130456 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T12:41:57Z |
| publishDate | 2021 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1304562022-10-01T10:33:29Z Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation Amini, Alexander A Gilitschenski, Igor Phillips, Jacob Moseyko, Julia Banerjee, Rohan Karaman, Sertac Rus, Daniela L Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology. Laboratory for Information and Decision Systems Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science In this work, we present a data-driven simulation and training engine capable of learning end-to-end autonomous vehicle control policies using only sparse rewards. By leveraging real, human-collected trajectories through an environment, we render novel training data that allows virtual agents to drive along a continuum of new local trajectories consistent with the road appearance and semantics, each with a different view of the scene. We demonstrate the ability of policies learned within our simulator to generalize to and navigate in previously unseen real-world roads, without access to any human control labels during training. Our results validate the learned policy onboard a full-scale autonomous vehicle, including in previously un-encountered scenarios, such as new roads and novel, complex, near-crash situations. Our methods are scalable, leverage reinforcement learning, and apply broadly to situations requiring effective perception and robust operation in the physical world. 2021-04-12T19:10:37Z 2021-04-12T19:10:37Z 2020-01 2021-04-07T12:14:16Z Article http://purl.org/eprint/type/JournalArticle 2377-3766 2377-3774 https://hdl.handle.net/1721.1/130456 Amini, Alexander et al. "Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation." IEEE Robotics and Automation Letters (April 2020): 5, 2 (January 2020): 1143 - 1150. en http://dx.doi.org/10.1109/lra.2020.2966414 IEEE Robotics and Automation Letters Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) MIT web domain |
| spellingShingle | Amini, Alexander A Gilitschenski, Igor Phillips, Jacob Moseyko, Julia Banerjee, Rohan Karaman, Sertac Rus, Daniela L Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation |
| title | Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation |
| title_full | Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation |
| title_fullStr | Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation |
| title_full_unstemmed | Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation |
| title_short | Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation |
| title_sort | learning robust control policies for end to end autonomous driving from data driven simulation |
| url | https://hdl.handle.net/1721.1/130456 |
| work_keys_str_mv | AT aminialexandera learningrobustcontrolpoliciesforendtoendautonomousdrivingfromdatadrivensimulation AT gilitschenskiigor learningrobustcontrolpoliciesforendtoendautonomousdrivingfromdatadrivensimulation AT phillipsjacob learningrobustcontrolpoliciesforendtoendautonomousdrivingfromdatadrivensimulation AT moseykojulia learningrobustcontrolpoliciesforendtoendautonomousdrivingfromdatadrivensimulation AT banerjeerohan learningrobustcontrolpoliciesforendtoendautonomousdrivingfromdatadrivensimulation AT karamansertac learningrobustcontrolpoliciesforendtoendautonomousdrivingfromdatadrivensimulation AT rusdanielal learningrobustcontrolpoliciesforendtoendautonomousdrivingfromdatadrivensimulation |