Generalizing Over Uncertain Dynamics for Online Trajectory Generation
We present an algorithm which learns an online trajectory generator that can generalize over varying and uncertain dynamics. When the dynamics is certain,the algorithm generalizes across model parameters. When the dynamics is partially observable, the algorithm generalizes across different observati...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Springer Nature
2021
|
| Online Access: | https://hdl.handle.net/1721.1/137619 |
| _version_ | 1811096557147127808 |
|---|---|
| author | Kim, Beomjoon Kim, Albert Dai, Hongkai Kaelbling, Leslie Lozano-Perez, Tomas |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Kim, Beomjoon Kim, Albert Dai, Hongkai Kaelbling, Leslie Lozano-Perez, Tomas |
| author_sort | Kim, Beomjoon |
| collection | MIT |
| description | We present an algorithm which learns an online trajectory generator that can generalize over varying and uncertain dynamics. When the dynamics is certain,the algorithm generalizes across model parameters. When the dynamics is partially observable, the algorithm generalizes across different observations. To do this, we employ recent advances in supervised imitation learning to learn a trajectory generator from a set of example trajectories computed by a trajectory optimizer. In experiments in two simulated domains, it finds solutions that are nearly as good as, and sometimes better than, those obtained by calling the trajectory optimizer online. The online execution time is dramatically decreased, and the off-line training time is reasonable. |
| first_indexed | 2024-09-23T16:45:34Z |
| format | Article |
| id | mit-1721.1/137619 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T16:45:34Z |
| publishDate | 2021 |
| publisher | Springer Nature |
| record_format | dspace |
| spelling | mit-1721.1/1376192023-01-10T16:32:25Z Generalizing Over Uncertain Dynamics for Online Trajectory Generation Kim, Beomjoon Kim, Albert Dai, Hongkai Kaelbling, Leslie Lozano-Perez, Tomas Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory We present an algorithm which learns an online trajectory generator that can generalize over varying and uncertain dynamics. When the dynamics is certain,the algorithm generalizes across model parameters. When the dynamics is partially observable, the algorithm generalizes across different observations. To do this, we employ recent advances in supervised imitation learning to learn a trajectory generator from a set of example trajectories computed by a trajectory optimizer. In experiments in two simulated domains, it finds solutions that are nearly as good as, and sometimes better than, those obtained by calling the trajectory optimizer online. The online execution time is dramatically decreased, and the off-line training time is reasonable. 2021-11-05T20:35:48Z 2021-11-05T20:35:48Z 2017-07 2019-06-04T15:21:12Z Article http://purl.org/eprint/type/ConferencePaper 2511-1256 2511-1264 https://hdl.handle.net/1721.1/137619 Kim, Beomjoon, Kim, Albert, Dai, Hongkai, Kaelbling, Leslie and Lozano-Perez, Tomas. 2017. "Generalizing Over Uncertain Dynamics for Online Trajectory Generation." en 10.1007/978-3-319-60916-4_3 Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Springer Nature MIT web domain |
| spellingShingle | Kim, Beomjoon Kim, Albert Dai, Hongkai Kaelbling, Leslie Lozano-Perez, Tomas Generalizing Over Uncertain Dynamics for Online Trajectory Generation |
| title | Generalizing Over Uncertain Dynamics for Online Trajectory Generation |
| title_full | Generalizing Over Uncertain Dynamics for Online Trajectory Generation |
| title_fullStr | Generalizing Over Uncertain Dynamics for Online Trajectory Generation |
| title_full_unstemmed | Generalizing Over Uncertain Dynamics for Online Trajectory Generation |
| title_short | Generalizing Over Uncertain Dynamics for Online Trajectory Generation |
| title_sort | generalizing over uncertain dynamics for online trajectory generation |
| url | https://hdl.handle.net/1721.1/137619 |
| work_keys_str_mv | AT kimbeomjoon generalizingoveruncertaindynamicsforonlinetrajectorygeneration AT kimalbert generalizingoveruncertaindynamicsforonlinetrajectorygeneration AT daihongkai generalizingoveruncertaindynamicsforonlinetrajectorygeneration AT kaelblingleslie generalizingoveruncertaindynamicsforonlinetrajectorygeneration AT lozanopereztomas generalizingoveruncertaindynamicsforonlinetrajectorygeneration |