Towards the development of a robust model for estimating wrist torque at different wrist angles

Several robotic devices have been developed to assist patients regain their autonomy. These robotic devices augment forces exerted by the device wearer. A majority of these devices are controlled by surface electromyography (SEMG) signals acquired from wearer´s muscles. Several regression models are available for estimation of wrist torques using SEMG signals. Common issues related to torque estimation models are degradation of model accuracy when the limb posture is altered. This degradation renders these models unsuitable for practical applications. For this reason, a model is sought that is unaffected by varying arm postures and joint angles. Obtained results for overcoming this problem are presented.