Robotics-assisted mirror rehabilitation therapy: A therapist-in-the-loop assist-as-needed architecture

This paper presents a therapist-in-the-loop framework for robotics-assisted mirror rehabilitation integrated with adaptive assist-as-needed therapy (ANT) that is adjusted based on the impairment level of the patient's affected limb. The framework, which is designed for patients with hemiparesis and/or hemispatial neglect, uses a patient's functional limb as the medium to transfer therapeutic training from the therapist to the patient's impaired limb (PIL). This allows the patient to use his/her functional limb to adjust the desired trajectory generated by the therapist if the trajectory is painful or uncomfortable for the PIL. In order to realize the adaptive patient-targeted therapy, two motor-function assessment metrics, performance symmetry and level of guidance, are proposed, providing real time, task-independent, and objective assessment of the PIL's motor deficiency. An adaptation law is also presented to adjust the intensity of the therapy delivered to the patient in real time and based on the aforementioned estimation of the impairment level of the PIL. Closed-loop system stability has been investigated in the presence of communication delays to facilitate tele/in-home rehabilitation. For this purpose, a combination of the Circle Criterion and the Small-Gain theorem has been applied to account both for communication time delays and the time-varying adaptive ANT. Results of experiments to investigate the performance of the proposed framework are reported.