Mohammad Mahdi Rusta - Department of Mechanical Engineering, Shiraz University, Shiraz, ۷۱۹۳۶-۱۶۵۴۸, Iran
Seyyed Arash Haghpanah - Department of Mechanical Engineering, Shiraz University, Shiraz, ۷۱۹۳۶-۱۶۵۴۸, Iran
Sajjad Taghvaei - Department of Mechanical Engineering, Shiraz University, Shiraz, ۷۱۹۳۶-۱۶۵۴۸, Iran

Manual wheelchair users rely on their upper limbs for independence and daily activities. The high incidence of upper limb injuries can be attributed to the significant muscular demands imposed by propulsion as a repetitive movement. People with spinal cord injury are at high risk for upper limb injuries, including neuromusculoskeletal pathologies and nociceptive pain, as human upper limbs are poorly designed to facilitate chronic weight-bearing activities, such as manual wheelchair propulsion. Comprehending the underlying biomechanical mechanisms of motor control and developing appropriate rehabilitation tasks are essential to deal with the effects of poor motor control on the performance of manual wheelchair users and prevent long-term upper limb disability, which can interrupt electrical signals between the brain and muscles. Functional electrical stimulation utilizes low-intensity electrical signals to artificially generate body movements by stimulating the damaged peripheral nerves of patients with impaired motor control. Therefore, this study investigates the central nervous system strategy to control human movements, which can be used for task-specific functional electrical stimulation rehabilitation therapy. To this aim, two degrees of freedom musculoskeletal model of the upper limb, including six muscles, is developed, and an optimal controller consisting of two separate optimal parts is proposed to track the desired trajectories in the joint space and estimate the optimal muscle activations regarding physiological constraints. The simulation results are validated with electromyography datasets collected from twelve participants. This study's primary advantages are generating optimal joint torques, accurate trajectory tracking, and good similarities between estimated and measured muscle activations.

Musculoskeletal modeling, Nonlinear model predictive controller, Static optimization, Manual wheelchair propulsion, Functional electrical stimulation