Haptic Teleoperation Based Rehabilitation Systems for Task-Oriented Therapy

摘要

This thesis focuses on the analysis and implementation of haptic teleoperation systems for home-based remote rehabilitation therapies. The main objective is to link the hand of a hospital-based therapist to the hand of a home-based disabled patient haptically, in order to simulate conventional in-hospital therapies. A new telerehabilitation paradigm involving telerobotic systems is proposed. The proposed ``Learn-and-replay" paradigm for task-oriented therapy consists of two phases: a therapist-in-loop phase where the therapist interacts with the patient through the haptic telerobotic interface to perform one or more repetitions of a cooperative therapy task, and a therapist-out-of-loop phase where the therapist's cooperative role in the task is played by the patient-side robot in future repetitions. Various technical aspects of this paradigm are explored in this thesis, including the therapist's arm impedance estimation and emulation through impedance control. One degree-of-freedom and a two degree-of-freedom advanced cooperative manipulation tasks are tested for proof of concept. Traditionally, human arm passivity is assumed for teleoperation system stability analysis. Recent research has shown that such an assumption can be inaccurate or too conservative. Especially, human arm can demonstrate active behaviour during telerehabilitation scenarios. A series-shunt approach is proposed to take into account some a priori information about the activity or excessive passivity of the human operator in order to derive more exact stability criteria. Both theoretical derivation and experimental validation are carried out in this work.