Design and implementation of an automated human activity monitoring application for wearable devices

We present the design and implementation of a system that monitors physical activities of a human user wearing a portable device equipped with inertial sensors. This system uses inertial sensor data to automatically classify human activities using a trained classifier. The portable device is used both as a sensing as well as an application platform with user interface. The software application includes the classifier and a programmable graphical user interface that displays feedback on physical fitness related metrics such as activity type, duration, and the amount of energy expended while performing the action. The application also enables users to review historical data and set future goals. Processed activity information is stored and transmitted using a wireless connection to a remote web-based server, which can publish this feedback to the users on a secure website as needed. This application aims at improving user's awareness and lifestyle by providing on-demand feedback using real-time, reflective and motivational modes. In short, we propose a system that integrates sensing, computing, and user interaction into a wearable platform with the purpose of monitoring and reporting users' physical performance metrics with the goal of motivating physical fitness or active lifestyle without interfering with their day-to-day activities.