Design and Torque Control of a Modular Reconfigurable Variable Stiffness Exoskeleton for Improved Transparency

Exoskeletons play a huge role in increasing the strength of normal people and assisting people with weak mobility. High mechanical robustness, low output impedance, reconfigurable hardware structure, and accurate torque control are essential for exoskeletons to enable safe and transparent physical human–robot interaction. In this article, first, a reconfigurable variable stiffness actuator with passive nonlinear variable stiffness and active variable stiffness characteristics is proposed to improve the adaptability and transparency of the exoskeleton. Second, a modular reconfigurable variable stiffness exoskeleton (MRVS-EXO) is proposed. The MRVS-EXO accommodates various configurations through reconfigurable and modular design, addressing the diverse needs of users. Then, a hysteresis modeling method based on “virtual deformation” is proposed to enable online torque estimation. This method employs the linearization method of “virtual deformation” to model hysteresis as segmented straight lines and curves, reducing computational performance requirements. Utilizing the hysteresis model, a cascade torque control method based on hysteresis compensation is employed to achieve precise torque control. Finally, the performance of the torque estimation and control method is experimentally verified, and the transparency of the exoskeleton is tested by going upstairs, downstairs, and walking outdoors experiments on subjects. Experimental results show that the proposed MRVS-EXO design and the torque control method enhance the transparency of the exoskeleton.