Reinforcement Learning for Mixed Autonomy Intersections
We propose a model-free reinforcement learning method for controlling mixed autonomy traffic in simulated traffic networks with through-traffic-only two-way and four-way intersections. Our method utilizes multi-agent policy decomposition which allows decentralized control based on local observati...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023
|
| Online Access: | https://hdl.handle.net/1721.1/148680 |
| _version_ | 1811070339000565760 |
|---|---|
| author | Yan, Zhongxia Wu, Cathy |
| author2 | Massachusetts Institute of Technology. Department of Civil and Environmental Engineering |
| author_facet | Massachusetts Institute of Technology. Department of Civil and Environmental Engineering Yan, Zhongxia Wu, Cathy |
| author_sort | Yan, Zhongxia |
| collection | MIT |
| description | We propose a model-free reinforcement learning method for controlling mixed
autonomy traffic in simulated traffic networks with through-traffic-only
two-way and four-way intersections. Our method utilizes multi-agent policy
decomposition which allows decentralized control based on local observations
for an arbitrary number of controlled vehicles. We demonstrate that, even
without reward shaping, reinforcement learning learns to coordinate the
vehicles to exhibit traffic signal-like behaviors, achieving near-optimal
throughput with 33-50% controlled vehicles. With the help of multi-task
learning and transfer learning, we show that this behavior generalizes across
inflow rates and size of the traffic network. Our code, models, and videos of
results are available at
https://github.com/ZhongxiaYan/mixed_autonomy_intersections. |
| first_indexed | 2024-09-23T08:34:23Z |
| format | Article |
| id | mit-1721.1/148680 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T08:34:23Z |
| publishDate | 2023 |
| publisher | IEEE |
| record_format | dspace |
| spelling | mit-1721.1/1486802023-03-24T03:50:40Z Reinforcement Learning for Mixed Autonomy Intersections Yan, Zhongxia Wu, Cathy Massachusetts Institute of Technology. Department of Civil and Environmental Engineering We propose a model-free reinforcement learning method for controlling mixed autonomy traffic in simulated traffic networks with through-traffic-only two-way and four-way intersections. Our method utilizes multi-agent policy decomposition which allows decentralized control based on local observations for an arbitrary number of controlled vehicles. We demonstrate that, even without reward shaping, reinforcement learning learns to coordinate the vehicles to exhibit traffic signal-like behaviors, achieving near-optimal throughput with 33-50% controlled vehicles. With the help of multi-task learning and transfer learning, we show that this behavior generalizes across inflow rates and size of the traffic network. Our code, models, and videos of results are available at https://github.com/ZhongxiaYan/mixed_autonomy_intersections. 2023-03-23T16:44:30Z 2023-03-23T16:44:30Z 2021 2023-03-23T15:48:02Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/148680 Yan, Zhongxia and Wu, Cathy. 2021. "Reinforcement Learning for Mixed Autonomy Intersections." 2021 IEEE International Intelligent Transportation Systems Conference (ITSC). en 10.1109/ITSC48978.2021.9565000 2021 IEEE International Intelligent Transportation Systems Conference (ITSC) Creative Commons Attribution-Noncommercial-Share Alike https://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf IEEE arXiv |
| spellingShingle | Yan, Zhongxia Wu, Cathy Reinforcement Learning for Mixed Autonomy Intersections |
| title | Reinforcement Learning for Mixed Autonomy Intersections |
| title_full | Reinforcement Learning for Mixed Autonomy Intersections |
| title_fullStr | Reinforcement Learning for Mixed Autonomy Intersections |
| title_full_unstemmed | Reinforcement Learning for Mixed Autonomy Intersections |
| title_short | Reinforcement Learning for Mixed Autonomy Intersections |
| title_sort | reinforcement learning for mixed autonomy intersections |
| url | https://hdl.handle.net/1721.1/148680 |
| work_keys_str_mv | AT yanzhongxia reinforcementlearningformixedautonomyintersections AT wucathy reinforcementlearningformixedautonomyintersections |