Admittance-based assistive control of a hip exoskeleton in a novel closed-loop MATLAB-Moco co-simulation framework

In response to the growing prevalence of musculoskeletal disorders in contemporary lifestyles, wearable robotic technologies have emerged as promising solutions to enhance physical performance and support rehabilitation. This study uses a novel MATLAB-Moco co-simulation platform integrating OpenSim Moco into closed-loop control frameworks for developing and evaluating assistive control strategies for exoskeleton robots. The platform leverages OpenSim's musculoskeletal modeling capabilities and Moco's optimal control solver to simulate assistive control strategies effectively. As a case study, an admittance controller was implemented for a custom hip exoskeleton. The human joint torque information from Moco model is fed to an admittance controller in MATLAB, with a simple pendulum dynamic model of the leg, to dynamically adjust the reference trajectory for the exoskeleton. Subsequently, a sliding mode controller fine-tunes the robot's torque to track the desired trajectory generated by the admittance controller. During a complete gait cycle, the proposed method reduced the human hip joint torque by 36.23%, while the desired experimental trajectory is accurately tracked. This indicates that the robot effectively supports the movements while maintaining user control and comfort. Additionally, robustness tests reveal that with ±30% error in the mass parameter, the assistance ratio fluctuates by less than ±3%, highlighting the effectiveness and robustness of the admittance controller, with only a simple pendulum model. The assistive algorithm is also straightforward to design and implement, requiring only hip joint kinematics and an approximate estimate of the physical characteristics of the user.