Detection and Classification for Freezing of Gaits in Parkinson's Disease Using Inertial Measurement Units

Freezing of Gaits(FOG) is a phenomenon in Parkin- son's Disease characterised by a short period of inability to initiate a gait or by a series of short steps with tremors on lower limbs, which mostly occurs when initiating gaits and turning. For researchers in health therapy, it is important to identify the instances of freezing of gait (FoG), in order to study the gait of a person suffering with Parkinson's disease. A video for walking of Parkinson's patient must be recorded so that researchers can label the time and type of FOG by watching the video again. To reduce the manual work and the cost of equipments, such as a video feed data, we use low cost Inertial Measurement Unit(IMU) to automatically detect and classify these occurrences of FoG.We use the instances of zero-velocity updates to detect FoG instances and classify them accordingly. With a decision machine combined with support vector machine analyzing ZUPTs and yaw angle from the Kalman Filter, we successfully detected and classified the FOG in our experiments.