THE DEVELOPMENT OF A WEARABLE MOTION ANALYSIS SYSTEM Sexton A, McGibbon C, Wilson A, Hughes G, Hudgins B Institute of Biomedical Engineering, University of New Brunswick, Fredericton, New Brunswick, Canada INTRODUCTION Between 40,000 and 50,000 Canadians sustain a stroke each year, and approximately 16,000 do not survive. 1 As such, there are presently about 300,000 Canadians currently living with the debilitating effects of stroke. 1 With an annual cost exceeding $2.7 billion, with $27.5 thousand per year per patient for acute care costs, there is a considerable challenge for rehabilitation professionals to provide efficacious treatments to restore function and reduce disability caused by stroke. Motor dysfunction experienced by stroke survivors can be due to any number of different functional syndromes such as paresis, ataxia, apraxia, visuo-perceptual deficits, or deafferentation. 2 A common functional deficit experienced by many stroke survivors is severe arm paresis. This can be characterized in general as loss of elbow and shoulder mobility caused by spasticity of flexor muscles, resulting in severe difficulties in self-care, such as feeding and dressing, and inability to perform common functional tasks, such as grasping and moving objects. Recent evidence suggests that arm motor function at 1 month post-stroke is one of the biggest predictors of stroke recovery. 3 It is therefore imperative that close monitoring of arm function in the weeks following stroke be clinically viable. Unfortunately, high costs and other burdens associated with frequent visits to the clinic or rehabilitation hospital often prevent adequate temporal resolution of arm mobility recovery assessments. The project described in this paper seeks to develop a solution to this common clinical problem, by developing, testing and implementing a wireless, wearable motion sensor device for capturing arm (elbow and shoulder) motion at any desired temporal resolution, and with accuracy and reliability exceeding that of commonly used clinical instruments. DESCRIPTION Validation of the Institute of Biomedical Engineering (IBME) elbow sensor system was accomplished by comparing the joint angle measured by the sensor system with a precision potentiometer, which served as the gold standard angle measurement, and with a Vicon M-Cam motion analysis system, which provided the baseline performance criteria. The goal of this sensor design was to provide both accuracy and precision equal to, or better than, that of the clinically accepted Vicon system. In addition to assessing accuracy and precision, within-test and test-retest repeatability were also assessed. Prototype Arm A prototype arm was constructed to simulate human arm movement, allow for accurate measurement of joint angles, and for testing and validation of sensor mounting strategies. The prototype arm consisted of two solid aluminum cylinders acting as upper and lower arm segments linked together with a journal bearing at the elbow joint to simulate elbow flexion/extension. The upper arm segment was affixed to one side of a pair of rotating disk plates that were linked via a thrust bearing to allow for shoulder abduction/adduction. As well, an L-bracket was bolted to the back of the second plate allowing for the entire device to be clamped onto a rigid supporting frame. Several aluminum collars were slid onto the upper and lower cylinders and adjusted in position with set screws to allow for the mounting and adjustment of sensors and Vicon reflective marker mounts. At the two joints, precision potentiometers were installed inside the sleeves of the bearings, providing analog signals of the joint angles. Motion Sensors The flexible ribbon portion of the ShapeTape sensor was affixed to the prototype arm’s upper and lower segments using plastic sliders or flexible rubber sheathing held in place with tape. The ShapeTape is an array of fiber optic based, 3D bend and twist sensors which provide accurate position and orientation information along its length. 4 The casing of the ShapeTape was attached to the rigid support using Velcro. A MicroStrain 3DM-GX1 inertial sensor was then attached “piggy-back” to the ShapeTape casing, also with Velcro. The MicroStrain 3DM-GX1 utilizes triaxial gyroscopes to track dynamic orientation and triaxial DC accelerometers along with triaxial magnetometers to track static orientation. 5 Data Acquisition A single PC ran both the Vicon data capture software and the IBME sensor data acquisition software (Figure #1). The Vicon software has an external start/stop