Abstract
A transformative advancement in the diagnosis of upper limb (UL) spasticity is on the horizon, moving beyond traditional reliance on clinicians' tactile assessments. Researchers, affiliated with UNIST have developed a robotic technology capable of quantifying spasticity with unprecedented accuracy, promising to enhance diagnosis, personalized rehabilitation, and compensation standards.
Led by Professor Sang Hoon Kang in the Department of Mechanical Engineering at UNIST, the team has created a novel method to objectively measure spasticity by applying subtle forces to a patient's arm and quantifying the resulting movement responses. Designed for rapid, quantitative assessment even by non-experts, this technology could significantly improve tailored rehabilitation strategies and support the establishment of standardized assessment criteria.
The team validated their approach using a 2-degree-of-freedom (2-DOF) direct-drive robotic system. Their findings revealed that even in such direct-drive configurations, small yet significant joint frictions remained, affecting measurement accuracy. This suggests that even well-known rehabilitation robots, like MIT-Manus-which are reputed for minimal friction-may still have residual issues that can compromise measurement reliability.
Furthermore, through experimental investigation, the researchers identified that these residual frictions are a primary cause of the nonlinear responses observed in human limbs during assessments. Prior studies attributed low measurement reliability to the nonlinear properties of the human arm, such as muscle hysteresis; however, this study demonstrates that the nonlinearities originate predominantly from the robotic system itself.
To address this, the team implemented an Internal Model Based Impedance Control (IMBIC) strategy, which nearly fully compensated for the robot's residual nonlinear friction. As a result, the robotic system's movements exhibited linear behavior, markedly increasing the trustworthiness of the measurements.
Seongil Hwang, first author of the study, stated, "Previous robotic methods for measuring spasticity faced reliability and nonlinearity issues largely due to the robot's internal friction rather than the patient's limb. Our findings show that by correcting for this friction, we can dramatically improve measurement accuracy and reliability, enabling more precise assessments of spasticity."
Spasticity, a common motor impairment following stroke or nerve injury, has traditionally been assessed manually-clinicians physically manipulate the patient's limb and rely on tactile sensation to evaluate severity. This approach is subjective, prone to variability depending on the examiner's skill, and challenging to quantify consistently across joint and movement directions.
Professor Kang emphasized, "Quantitative, objective measurement of spasticity will aid in designing more effective rehabilitation protocols and establishing standardized evaluation metrics. We are also planning to collaborate with UNIST's upcoming specialized public hospital for occupational and physical medicine, opening pathways for real-world clinical application." Kang is also a concurrent professor at the University of Maryland School of Medicine in the United States.
This research was conducted with the participation of Dr. Hyunah Kang as a co-author. It was supported by the Ministry of Science and ICT (MSIT) and the National Research Foundation of Korea (NRF). The study was published in IEEE Transactions on Neural Systems and Rehabilitation Engineering, a leading journal in the field of rehabilitation medicine on March 26, 2025.
Journal Reference
Seongil Hwang, Hyunah Kang, and Sang Hoon Kang, "Assessing Linearity in Multi-Joint Upper Limb Dynamics Under Small Perturbations for Reliable Mechanical Impedance Estimation," IEEE TNSRE, (2025).
