| Summary: | Assisting human locomotion with a wearable robotic orthosis is still quite challenging, largely due to the complexity of the neuromusculoskeletal system, the time-varying dynamics that accompany motor adaptation, and the uniqueness of every individual’s response to the assistance given by the robot. To this day, these devices have not met their well-known promise yet, mostly due to the fact that they are not perfectly suitable for the rehabilitation of neuropathologic patients. One of the main challenges hampering this goal still relies on the interface and co-dependency between the human and the machine. Nowadays, most commercial exoskeletons replay pre-defined gait patterns, whereas research exoskeletons are switching to controllers based on optimized torque profiles. In most cases, the dynamics of the human musculoskeletal system are still ignored and do not take into account the optimal conditions for inducing a positive modulation of neuromuscular activity. This is because both rehabilitation strategies are still emphasized on the macro level of the whole joint instead of focusing on the muscles’ dynamics and activity, which are the actual anatomical elements that may need to be rehabilitated. Strategies to keep the human in the loop of the exoskeleton’s control laws in real-time may help to overcome these challenges. The main purpose of the present dissertation is to make a paradigm shift in the approach on how the assistance that is given to a subject by an exoskeleton is modelled and controlled during physical rehabilitation. Therefore, in the scope of the present work, it was intended to design, concede, implement, and validate a real-time muscle-in-the-loop optimization model to find the best assistive support ratio that would induce optimal rehabilitation conditions to a specific group of impaired muscles while having a minimum impact on the other healthy muscles. The developed optimization model was implemented in the form of a plugin and was integrated on a neuromechanical model-based interface for driving a bilateral ankle exoskeleton. Experimental pilot tests evaluated the feasibility and effectiveness of the model. Results of the most significant pilots achieved EMG reductions up to 61 ± 3 % in Soleus and 41 ± 10 % in Gastrocnemius Lateralis. Moreover, results also demonstrated the efficiency of the optimization’s specific reduction on rehabilitation by looking into the muscular fatigue after each experiment. Finally, two parallel preliminary studies emerged from the pilots, which looked at muscle adaptation, after a new assistive condition had been applied, over time and at the effect of the lateral positioning of the exoskeleton’s actuators on the leg muscles.
|
|---|