Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2403.04007 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913256592900096 |
|---|---|
| author | Suttle, Wesley A. Sharma, Vipul K. Kosaraju, Krishna C. Sivaranjani, S. Liu, Ji Gupta, Vijay Sadler, Brian M. |
| author_facet | Suttle, Wesley A. Sharma, Vipul K. Kosaraju, Krishna C. Sivaranjani, S. Liu, Ji Gupta, Vijay Sadler, Brian M. |
| contents | We develop provably safe and convergent reinforcement learning (RL) algorithms for control of nonlinear dynamical systems, bridging the gap between the hard safety guarantees of control theory and the convergence guarantees of RL theory. Recent advances at the intersection of control and RL follow a two-stage, safety filter approach to enforcing hard safety constraints: model-free RL is used to learn a potentially unsafe controller, whose actions are projected onto safe sets prescribed, for example, by a control barrier function. Though safe, such approaches lose any convergence guarantees enjoyed by the underlying RL methods. In this paper, we develop a single-stage, sampling-based approach to hard constraint satisfaction that learns RL controllers enjoying classical convergence guarantees while satisfying hard safety constraints throughout training and deployment. We validate the efficacy of our approach in simulation, including safe control of a quadcopter in a challenging obstacle avoidance problem, and demonstrate that it outperforms existing benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_04007 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Sampling-based Safe Reinforcement Learning for Nonlinear Dynamical Systems Suttle, Wesley A. Sharma, Vipul K. Kosaraju, Krishna C. Sivaranjani, S. Liu, Ji Gupta, Vijay Sadler, Brian M. Machine Learning Optimization and Control We develop provably safe and convergent reinforcement learning (RL) algorithms for control of nonlinear dynamical systems, bridging the gap between the hard safety guarantees of control theory and the convergence guarantees of RL theory. Recent advances at the intersection of control and RL follow a two-stage, safety filter approach to enforcing hard safety constraints: model-free RL is used to learn a potentially unsafe controller, whose actions are projected onto safe sets prescribed, for example, by a control barrier function. Though safe, such approaches lose any convergence guarantees enjoyed by the underlying RL methods. In this paper, we develop a single-stage, sampling-based approach to hard constraint satisfaction that learns RL controllers enjoying classical convergence guarantees while satisfying hard safety constraints throughout training and deployment. We validate the efficacy of our approach in simulation, including safe control of a quadcopter in a challenging obstacle avoidance problem, and demonstrate that it outperforms existing benchmarks. |
| title | Sampling-based Safe Reinforcement Learning for Nonlinear Dynamical Systems |
| topic | Machine Learning Optimization and Control |
| url | https://arxiv.org/abs/2403.04007 |