[home] [Personal Program] [Help]

---

Session: Rehabilitation
Session starts: Friday 27 January, 14:00
Presentation starts: 15:00
Room: Room 530

---

Suzanne Filius (Delft University of Technology)
Mariska Janssen (Radboud University Medical Center; Klimmendaal)
Herman van der Kooij (University of Twente)
Jaap Harlaar (Delft University of Technology; Erasmus MC University Medical Center)

Abstract:
There is a need for active arm supports in people with progressive muscle weakness such as Duchenne Muscular Dystrophy (DMD) to support functional arm movements during daily activities to enhance their independence, social participation and their quality of live. The best technical strategy to support functional arm movements remains challenging. This study serves as exploratory pilot study in the development of a motorized arm support for daily use in boys and men with DMD (Brooke Scale >3). This study aims to develop 4 different high level control strategies (e.g. modelled-based, hybrid-based, measured-based, and personalized-based) and compares the compensation efficacy and preferences in an active elbow support in healthy subjects. The study is performed in 12 healthy male participants and exists out of 2 sessions: 1) measurement of the passive interaction forces around the elbow joint in horizontal and vertical plane (relaxation task), 2) comparison of the four compensation strategies in vertical plane (position task). The passive interaction forces where measured with a force-torque sensor at the sleeve interface and the compensation support was provided by a motor at the elbow joint with a compensation gain of 80%. The modelled-based strategy makes use of a gravitational kinematic model, using an estimation of mass and center of mass (COM) of the forearm and hand. The measured-based strategy makes use of all passive forces measured in session 1 in the vertical plane. The hybrid-based strategy makes use of a combination of the kinematic model and the measured passive forces in the horizontal plane. Personalized-based makes use of a model of both gravity and joint impedance fit to the measured passive forces in vertical plane. The compensation efficacy was determined by the area under the curve (AUC) of the surface electromyography (sEMG) of the m. biceps brachii and long head of m. triceps muscles, and the tracking error during the target position tracking task. Out of the 12 (aged 24 to 35 years), 4 preferred the modelled-based, 4 preferred measured-based, 3 preferred personalized-based and 1 the hybrid-based compensation strategy. The root mean square (RMS) values of the torque-angle profiles of the modelled-based and hybrid-based were 13% and 20% higher compared to both the measured-based and personalized-based compensation strategy. On average the EMG AUC of the biceps muscle was 34% lower in the modelled-based and 21% lower for the hybrid-based compared to the measured-based. In the triceps muscles and tracking error no significant differences are expected. The discrepancy between the measured-based and modelled-based is expected to be the result of the horizontal plane passive forces (e.g. elbow joint impedance) and errors in the parameter estimation. This difference is expected to be larger in DMD due to elevated muscle stiffness and deviating anthropometry. The effect on the compensation efficacy (performance and muscle activity) in healthy subjects is only small but is expected to be larger in DMD. This will be investigated in the future. This work is part of Wearable Robotics program (P16-05), (partly) funded by the Dutch Research Council (NWO), The Netherlands.