Predictive Closed-loop Control of Musculotendon Force: Towards a Framework for Human Locomotion

Exosuits and exoskeletons show promise in aiding locomotion for both healthy and pathological populations. However, the impact of exosuit assistance on the underlying muscle-tendon loading remains unclear. In this study, our objective is to gain control over the mechanics of biological tissues, offering avenues for preventing musculoskeletal injuries and tailoring rehabilitation treatments with unprecedented precision. To achieve this goal, we introduce a novel framework utilizing Nonlinear Model Predictive Control (NMPC) for closed-loop control of Achilles tendon force during simulated human ankle joint motion with parallel exoskeletal actuation. The proposed NMPC framework integrates a computationally efficient inner model consisting of explicit, closed-form Ordinary Differential Equations (ODEs) governing muscle-tendon dynamics and ankle joint behavior with parallel actuation. We assess the controller's performance through hopping simulations, varying muscle excitation amplitude and frequencies. Furthermore, we extend the model to enable the use of the same control framework for musculotendon unit (MTU) force control in walking scenarios. The results confirm the versatility of our proposed model across different muscles and gait patterns, demonstrating its suitability for real-time applications due to its low computational time.