Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.19945 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914403412082688 |
|---|---|
| author | Zhang, Zhouyu Chiu, Chih-Yuan Chou, Glen |
| author_facet | Zhang, Zhouyu Chiu, Chih-Yuan Chou, Glen |
| contents | We present an inverse dynamic game-based algorithm to learn parametric constraints from a given dataset of local Nash equilibrium interactions between multiple agents. Specifically, we introduce mixed-integer linear programs (MILP) encoding the Karush-Kuhn-Tucker (KKT) conditions of the interacting agents, which recover constraints consistent with the local Nash stationarity of the interaction demonstrations. We establish theoretical guarantees that our method learns inner approximations of the true safe and unsafe sets. We also use the interaction constraints recovered by our method to design motion plans that robustly satisfy the underlying constraints. Across simulations and hardware experiments, our methods accurately inferred constraints and designed safe interactive motion plans for various classes of constraints, both convex and non-convex, from interaction demonstrations of agents with nonlinear dynamics. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_19945 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Constraint Learning in Multi-Agent Dynamic Games from Demonstrations of Local Nash Interactions Zhang, Zhouyu Chiu, Chih-Yuan Chou, Glen Machine Learning Systems and Control We present an inverse dynamic game-based algorithm to learn parametric constraints from a given dataset of local Nash equilibrium interactions between multiple agents. Specifically, we introduce mixed-integer linear programs (MILP) encoding the Karush-Kuhn-Tucker (KKT) conditions of the interacting agents, which recover constraints consistent with the local Nash stationarity of the interaction demonstrations. We establish theoretical guarantees that our method learns inner approximations of the true safe and unsafe sets. We also use the interaction constraints recovered by our method to design motion plans that robustly satisfy the underlying constraints. Across simulations and hardware experiments, our methods accurately inferred constraints and designed safe interactive motion plans for various classes of constraints, both convex and non-convex, from interaction demonstrations of agents with nonlinear dynamics. |
| title | Constraint Learning in Multi-Agent Dynamic Games from Demonstrations of Local Nash Interactions |
| topic | Machine Learning Systems and Control |
| url | https://arxiv.org/abs/2508.19945 |