Control System and Graphical User Interface Design of an Upper-Extremity Rehabilitation Robot

Abstract

Stroke is one of the leading causes of death, physical disability, and loss of brain functionality each year, especially amongst older adults. The ability to access good quality post-stroke rehabilitation exercises is essential for stroke survivors to maximize their potential to regain skills and physical abilities. Robot-assisted therapy is showing promise as a way to provide stroke survivors with engaging, challenging, and repetitive tasks while delivering measured therapy that is able to objectively evaluate patients’ progress. Among several challenges that are associated with the design of rehabilitation robots (e.g., the mechanical structure, the actuator types, the control strategies), the design of the control strategy is one of the most critical. Depending on the type of patient and the severity of the impairment of motor control, various control strategies could be applied for the recovery of the impaired limb in stroke survivors using robot-assisted therapy. Research is needed into the development of how best to control rehabilitation robots; this includes both the internal control algorithms and the User Interface (UI) for therapists.

As such, the first objective of this research is to design and implement a motion controller and force-field controller for a 2-Degree of Freedom (DOF) manipulandum upper-extremity rehabilitation robot that is able to deliver planar rehabilitation exercises for stroke survivors while taking therapeutic rehabilitation goals into account. The motion control algorithm can precisely follow a prescribed time-dependent trajectory whereas the force-control method will only provide assistance (or even resistance to introduce extra challenge) to the patient to do the task rather than forcing the movement. For doing the simulation studies, a motor control model of post-stroke patients was proposed. The effectiveness of these controllers was explored in simulations and it was observed that the developed force-field algorithm had a positive effect on the motor control recovery for a simulated patient. The simulation results also indicated that the resistive mode of therapy would result in better outcomes after the therapy which aligns with experimental studies by other researchers. In addition, a novel adaptive algorithm was proposed for fine-tuning the proposed force-field parameters based on the performance of the patient during the therapy as a subject specific controller can help to achieve a desirable performance for each patient. While this approach is promising, the effectiveness of the adaptation rule has yet to be evaluated on real patients in the future.

To enable effective access and use of the robot, the controller needs to be visualized through a Graphical User Interface (GUI) in a way that therapists can understand and use. The second goal of this thesis research was to work with therapists to collaboratively design an intuitive to use GUI for therapists to control the robot and provide objective information on patients’ performance. The identification of features and feedback on the intuitiveness of the GUI developed in this research highlights the value of collaborative design between engineers and therapists to create the interface that enables therapists to control the rehabilitation robot. This research also identifies the need for collaborative GUI design with patients as their needs and preference may be different from therapists. During the collaborative GUI design, it was observed that including obstacles and force-field method might be a possible useful method for supporting patients’ movement trajectory, not only because therapists can adjust the force strength to suit a specific patient, but also because they can use its numerical data for objective measurement of patients’ performance. Therapists who participated in this research stated that objective measurements (i.e., trajectory smoothness, speed, mobility range, and error) could be used to evaluate the patient performance. While rehabilitation robots are different in terms of mechanical structure, work-space, and the exercise that they can provide, similar methods could be used for supporting patients’ movement trajectory and performance evaluation. As the GUI is the first prototype, it needs to be used with and evaluated by therapists and patients to ascertain if the information presented in the GUI is intuitive and to explore if they can understand it or use it.