Développement d’un outil calculant en temps réel les moments de réaction articulaires aux épaules durant la propulsion en fauteuil roulant manuel sur simulateur par dynamique inverse

Abstract

Approximately 2.6 million Canadians aged 15 and older have difficulty moving in their daily lives. To help them in their mobility, various assistive technologies are available, depending on the individual's disability, including manual wheelchairs (MW). However, repetitive propulsion in MW can lead to musculoskeletal disorders, shoulder pain, and worsen the user's disability. Estimating the shoulder joint moment is an indicator used to prevent these injury risks. A previous study has already shown that training on a simulator can modify a user's propulsion technique to increase a real-time variable and make the participant more efficient during propulsion. However, it also increased the shoulder joint moment. In this study, the goal was to design a device to make the shoulder joint moment available in real-time (RT) using the inverse dynamics method, so that it could be used in the simulator training mechanism in the future. To achieve this, hardware and software upgrades were first made to the simulator, replacing the RT computer with a new specialized calculation computer. Subsequently, an optoelectronic system was installed to measure the upper limb kinematics of the user during propulsion on the simulator, and a communication protocol between the optoelectronic system and the RT computer was developed to read the transmitted RT signal. Next, the experimental protocol and inverse dynamics method used in a reference study were modified and adapted for realtime use. To validate the new RT method developed in this study, previous data from 18 spinal cord injured MW users who propelled on the simulator were used. The simulator, equipped with 2 instrumented wheels, measured forces and reaction moments, while an optoelectronic system measured upper limb kinematics. With these two measurements, the shoulder moments were calculated using the reference method, which was not in real-time. Then, in the present study, the saved inputs were transmitted at a frequency of 120 Hz to the calculation computer to simulate real-time data collection. The validation of this software was performed on the dependent variable corresponding to the average of the maximum and minimum peaks of the shoulder joint moment cycles along the three anatomical axes. A Bland-Altman analysis showed a difference between the two methods ranging from -0.3 to 1.1 Nm for adduction peak, -0.1 to 0.5 Nm for internal rotation, and -1.1 to 0.8 Nm for flexion in the calculated shoulder joint moment of the left shoulder. However, a Monte Carlo simulation of 1000 iterations on one subject, with variable noise representing the measurement noise inherent to the use of instrumented wheels, showed that the results obtained with the reference method had an uncertainty of ±1.4 Nm for adduction/abduction, ±0.6 Nm for internal/external rotation, and ±1.6 Nm for flexion/extension. By comparing these two analyses, it can be observed that the confidence intervals obtained from the Bland-Altman analysis are all within the tolerance intervals obtained from the Monte Carlo simulation. Therefore, the error associated with using one method compared to the other is small compared to the inherent error associated with the main measuring instrument, and the new real-time inverse dynamics method can be validated. The analysis for the right shoulder led to the same conclusion. In this study, real-time estimation of the shoulder joint moment was made available on the LIO simulator and can now be added to the training mechanism to limit shoulder load. Finally, the computational capacity of the new RT computer in the simulator could allow for the implementation of new RT measures such as muscle activity using electromyography sensors. The ultimate goal of such a tool would be to help MW users by predicting injury risks through their propulsion and modifying their propulsion technique through simulator training to minimize these risks.

Date	10 Jul 2023
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Rachid Aissaoui (Supervisor) & Felix Chenier (Co-supervisor)