Cost-Matching Model Predictive Control for Efficient Reinforcement Learning in Humanoid Locomotion
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911553476886528 |
|---|---|
| author | Cai, Wenqi Vamvoudakis, Kyriakos G. Gros, Sébastien Tzes, Anthony |
| author_facet | Cai, Wenqi Vamvoudakis, Kyriakos G. Gros, Sébastien Tzes, Anthony |
| contents | In this paper, we propose a cost-matching approach for optimal humanoid locomotion within a Model Predictive Control (MPC)-based Reinforcement Learning (RL) framework. A parameterized MPC formulation with centroidal dynamics is trained to approximate the action-value function obtained from high-fidelity closed-loop data. Specifically, the MPC cost-to-go is evaluated along recorded state-action trajectories, and the parameters are updated to minimize the discrepancy between MPC-predicted values and measured returns. This formulation enables efficient gradient-based learning while avoiding the computational burden of repeatedly solving the MPC problem during training. The proposed method is validated in simulation using a commercial humanoid platform. Results demonstrate improved locomotion performance and robustness to model mismatch and external disturbances compared with manually tuned baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_28243 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Cost-Matching Model Predictive Control for Efficient Reinforcement Learning in Humanoid Locomotion Cai, Wenqi Vamvoudakis, Kyriakos G. Gros, Sébastien Tzes, Anthony Robotics Systems and Control In this paper, we propose a cost-matching approach for optimal humanoid locomotion within a Model Predictive Control (MPC)-based Reinforcement Learning (RL) framework. A parameterized MPC formulation with centroidal dynamics is trained to approximate the action-value function obtained from high-fidelity closed-loop data. Specifically, the MPC cost-to-go is evaluated along recorded state-action trajectories, and the parameters are updated to minimize the discrepancy between MPC-predicted values and measured returns. This formulation enables efficient gradient-based learning while avoiding the computational burden of repeatedly solving the MPC problem during training. The proposed method is validated in simulation using a commercial humanoid platform. Results demonstrate improved locomotion performance and robustness to model mismatch and external disturbances compared with manually tuned baselines. |
| title | Cost-Matching Model Predictive Control for Efficient Reinforcement Learning in Humanoid Locomotion |
| topic | Robotics Systems and Control |
| url | https://arxiv.org/abs/2603.28243 |