Passive Robotics for Accelerated Collaborative Therapy

Each year 110,000 stroke cases are reported in the UK. 300,000 people are living with post-stroke complications, which for many include upper limb impairment. After-stroke therapy is most effective during the first 12 weeks, but improvement can continue for months or even years. To date, a number of devices prioritising early stroke rehabilitation have been developed, however most of these devices are active devices and cannot be used by patients without constant supervision due to safety concerns. In this work a low-cost, portable inherently safe rehabilitation robot is introduced. The robot is designed for use on existing table space in a home environment. Estimates of the robot's 2D position and orientation are computed by fusing data from two tracking systems, each utilizing a different sensor type: laser optical sensors and a webcam. Two laser optical sensors are mounted on the underside of the robot and track the relative motion of the robot with respect to the surface on which it is placed. The webcam is positioned directly above the workspace, mounted on a fixed stand, and tracks the robot's position with respect to a fixed coordinate system. The optical sensors sample the position data at a higher frequency than the webcam. A position and orientation fusion scheme is proposed to fuse the data from the two tracking systems. Active movements of a patient using the robot are assisted in real time by a custom-designed inherently safe guidance mechanism based on a COBOT unicycle architecture and capable of creating 2D virtual constraints. The guidance mechanism is mounted on the underside of the robot and in its centre, and is accompanied by four custom designed caster wheels with low rolling resistance. The guidance mechanism is actuated with a stepper motor, but no active force acting on the upper limb of the patient is generated. A prototype of the robot was manufactured and evaluated in a series of experiments including user testing with 38 healthy bodied adults. It was found that the robot is clearly capable of applying a passive force of suffcient magnitude and in the right direction to increase the accuracy with which able bodied adults can perform reaching arm movements. The resistance to movement of the robot, both friction and inertia are suffciently low that they do not appear to reduce the subjects' ability to move at normal reaching movement speeds.