Enforcing Constraints over Learned Policies via Nonlinear MPC: Application to the Pendubot
| Autor: | Leonardo Lanari, Massimo Cefalo, Giulio Turrisi, Giuseppe Oriolo, Valerio Modugno, B. Barros Carlos |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
0209 industrial biotechnology
Computer science Control (management) 02 engineering and technology Time step NMPC 020901 industrial engineering & automation Control theory 0202 electrical engineering electronic engineering information engineering Reinforcement learning Underactuated robots business.industry 020208 electrical & electronic engineering Work (physics) Model based control Learning control Robotics Optimal control Nonlinear system Model predictive control Control and Systems Engineering Nonlinear model Artificial intelligence business |
| Zdroj: | IFAC-PapersOnLine. 53:9502-9507 |
| ISSN: | 2405-8963 |
| Popis: | In recent years Reinforcement Learning (RL) has achieved remarkable results. Nonetheless RL algorithms prove to be unsuccessful in robotics applications where constraints satisfaction is involved, e.g. for safety. In this work we propose a control algorithm that allows to enforce constraints over a learned control policy. Hence we combine Nonlinear Model Predictive Control (NMPC) with control-state trajectories generated from the learned policy at each time step. We prove the effectiveness of our method on the Pendubot, a challenging underactuated robot. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|