| Author | Harischandra, Patikiri Arachchige Diluka |
| Note | A thesis submitted in partial fulfillment of the requirements for the
degree of Master of Engineering in
Mechatronics |
| Publisher | Asian Institute of Technology |
| Abstract | Conventional rehabilitation methods are tiresome for the therapists and also inefficient. The amount of patients a therapist must handle is quite high and the feasible solution for this is robot aided therapy. In this thesis, improvements for robot aided rehabilitation in three research areas, namely, parameter estimation of human arm, tele-rehabilitation, and bimanual rehabilitation are presented. In most of the studies, the human arm inertia is considered as a system disturbance in rehabilitation robots. Human limbs are also not modelled as individual anthropometry vary from patient to patient. The best option is to estimate the arm inertia and gravity torque to the best accuracy with minimum time. Estimated parameters could be effectively used in the controller loop to achieve precise motion control. A novel Reaction Torque Observer (RTOB) based estimation technique which is updated in real-time using learning and recursive algorithms is presented. The applicability of the proposed methods are verified using a real world hardware system. The accuracy of the estimated parameters suggests that the proposed methods are suitable for implementation in rehabilitation robots. Operationality of teleoperated systems can be improved by gravity compensation. This thesis proposes an online gravity estimation method which could be used in a tele-rehabilitation system. Gravity torque is estimated and compensated in the system such that the therapist would feel only the abnormalities of the patient’s arm. Online recursive numerical parameter estimation method was proposed to be used with the reaction torque observer’s output data. The proposed method was practically verified with a master slave telerehabilitation system. This thesis further proposes bimanual rehabilitation robot system for the improvement of inter-limb coordination. Existing rehabilitation machines which are designed for resistance training do not follow profile of the human arm torque curve therefore the workout is inefficient. Ideally, a rehabilitation robot should provide both high impedance and assistance when the patient is unable to work against the resistance. This thesis proposes a novel impedance controlled bimanual robot using the reaction torque observer with seamless adjustment of resistance. The proposed controller is tested with two options. Results show the applicability of the proposed methods. |
| Year | 2017 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Department of Industrial Systems Engineering (DISE) |
| Academic Program/FoS | Microelectronics (ME) |
| Chairperson(s) | Abeykoon, A. M. Harsha S. |
| Examination Committee(s) | Bohez, Erik L. J.;Manukid Parnichkun; |
| Scholarship Donor(s) | AIT Fellowship; |
| Degree | Thesis (M.Eng.) -- Asian Institute of Technology, 2017 |