Gait Analysis with IMU - Gaining New Orientation Information of the Lower Leg.

In this paper we present an application for the analysis and characterisation of gait motion. Using motion data from Inertial Measurement Units (IMUs), seven relevant parameters are measured that extensively charaterize the gait of individuals. Our application uses raw and processed IMU data, where the processed data is the result of filtering the IMU data with a Magdwick filter. The filtered data offers orientation information and is relatively drift free. The IMU data is used to train a three layer neural network that can then extract individual footsteps from an IMU dataset. Results with different test persons show that our application can successfully characterize gait motion on an individual basis and can serve for the clinical assesment and evaluation of abnormal or pathological gait.