What question did this study set out to answer?

To develop an online model-based reinforcement learning algorithm for real-world robotic control.

Efficient model-based reinforcement learning for robot control via online optimization

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial shows improved control policy efficiency in robotic systems, suggesting reliable on-robot learning.

To develop an online model-based reinforcement learning algorithm for real-world robotic control.
Built a dynamics model from real-time interaction data for policy updates.
Used online optimization analysis to derive sublinear regret bounds.
Conducted experiments on a hydraulic excavator arm and a soft robot arm.
The algorithm achieved strong sample efficiency, matching model-free methods within hours.
Demonstrated robust adaptation to random changes in payload conditions.

Fang et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial shows improved control policy efficiency in robotic systems, suggesting reliable on-robot learning.

To develop an online model-based reinforcement learning algorithm for real-world robotic control.
Built a dynamics model from real-time interaction data for policy updates.
Used online optimization analysis to derive sublinear regret bounds.
Conducted experiments on a hydraulic excavator arm and a soft robot arm.
The algorithm achieved strong sample efficiency, matching model-free methods within hours.
Demonstrated robust adaptation to random changes in payload conditions.

Fang et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark