Adaptive Hybrid Compliant Control of a Desktop Upper-Limb Rehabilitation Robot

• Auwalu M. ABDULLAHI

  Author

• Musa M. BELLO

  Author

• Ado HARUNA

  Author

• Amina I. KHALEEL

  Author

• Lubabatu B. ILA

  Author

• Abdurrahaman M. BELLO

  Author

Robotic rehabilitation systems, including exoskeletons and endpoint robots, are increasingly employed in physiotherapy for patients recovering from spinal cord injuries and stroke. Conventional rehabilitation relies heavily on physiotherapists and hospital-based sessions, which often involve scheduling constraints and long waiting times due to the high number of patients. Robotic assistance can alleviate this burden by enabling more frequent and consistent therapy while maintaining safe and effective human–robot interaction. Achieving safe interaction requires compliant control strategies capable of regulating the contact force between the patient and the robot, particularly in rehabilitation scenarios involving uncertain environments where interaction dynamics vary depending on soft or stiff contact conditions. Hybrid impedance and admittance position control (HIPC) has been widely adopted for this purpose. However, conventional HIPC typically employs fixed impedance parameters, which may lead to excessive energy consumption and reduced adaptability during therapy. This paper proposes a Hybrid Adaptive Impedance and Position Control (HAIPC) framework for robot-assisted rehabilitation. The proposed approach integrates a genetic algorithm–tuned PD-based admittance position controller for trajectory guidance with an adaptive impedance controller that dynamically updates stiffness and damping gains. An extended state observer is employed to estimate the contact torque, enabling adaptive control without the need for torque/force sensors. Simulation results demonstrated that the proposed HAIPC significantly reduces energy consumption by about 64% and reduces the contact torque tracking error to a similar extent compared to the conventional HIPC. The proposed strategy improves adaptability during human–robot interaction while offering a cost-effective sensorless implementation for rehabilitation robots.

• FJET_21_48_48

Copyright (c) 2026 FUDMA Journal of Engineering and Technology

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.