Related Experiment Video
Updated: May 14, 2026
06:58
A Structured Rehabilitation Protocol for Improved Multifunctional Prosthetic Control: A Case Study
Published on: November 6, 2015
Model Predictive Control-Based Assist-as-Needed Strategy for Reducing Motor Slacking in Robot-Assisted Rehabilitation
Choonggun Kim1, Youngjin Moon2, Jaesoon Choi2
1Department of Mechanical Engineering, Sogang University, Seoul 04107, Republic of Korea.
Sensors (Basel, Switzerland)
|May 13, 2026
Summary
This study introduces a new Assist-as-Needed (AAN) strategy for upper-limb rehabilitation robots that prevents motor slacking by optimizing robotic assistance and user participation. This model predictive control (MPC) approach enhances rehabilitation robot effectiveness.
Area of Science:
- Robotics
- Rehabilitation Engineering
- Control Systems
Background:
- Conventional Assist-as-Needed (AAN) strategies in upper-limb rehabilitation robots often mask reduced user effort due to reliance on a single error-based coefficient.
- This lack of explicit user participation monitoring can lead to motor slacking, hindering effective rehabilitation.
Purpose of the Study:
- To propose and validate a novel model predictive control (MPC)-based AAN strategy for upper-limb rehabilitation robots.
- To specifically address and mitigate motor slacking by decoupling assistance and user participation.
- To enhance the effectiveness of robot-assisted rehabilitation by promoting genuine user engagement.
Main Methods:
- Developed a two-channel admittance structure with jointly optimized robotic-assistance gain (Ak) and user-participation-reflection gain (Bk) within a convex MPC formulation.
- Incorporated a cost function addressing trajectory tracking, participation-aware force alignment, assistance suppression, and passivity via energy-tank constraints.
- Validated the controller through two experiments on a mobile upper-limb rehabilitation robot, including comparisons with baseline AAN methods under induced motor-slacking conditions.
Main Results:
- The proposed controller demonstrated differential adaptation of Ak and Bk across varying instructed contribution levels, with participation ratios ranging from 0.103 to 0.879.
- In motor-slacking conditions, the proposed controller achieved a significantly lower aggregate human-contribution ratio (0.282) compared to error-based (0.595) and forgetting-factor (0.535) baselines.
- The controller explicitly represented externally imposed reductions in participation, unlike baseline methods where reductions were masked by assistive compensation.
Conclusions:
- The proposed MPC-based AAN strategy effectively mitigates motor slacking in upper-limb rehabilitation robots.
- This approach enhances participation-preserving and anti-slacking capabilities in robot-assisted rehabilitation.
- The findings suggest a more explicit and effective method for controlling robotic assistance based on user engagement.
