An application for IMU sensor data analysis. This is the algorithm portion of my capstone team project that I developed. It is currently in development to further improve the algorithm performance and the code structure.
In cross-country ski training, wearable IMU sensors are placed on athletes' poles and boots to capture movement data. These IMU sensors measure triaxial linear acceleration, triaxial angular velocity and triaxial magnetic force and record them as time-series data. Data analysts extract insights from this data, and coaches use these insights to provide feedback on athletes' skiing techniques. In many instances, identifying the ski steps taken by an athlete during a workout is an important first step in IMU data processing that opens the way to further analysis of skiing techniques. Thus, the goal for this application is to provide an easy way to identify the ski steps in a workout.
TODO: Image of full IMU data with workout data labeled. Image of workout data with steps labeled.
See here for instructions to set up the development environment locally.
See model for more details on how we automatically identify the ski steps.