Everyone has been involved with sports activities at some point, whether it is running, cycling, swimming, or any other activity. Once we finish the activity, it always comes in handy to see the summary of it (kilometers driven, calories burned, etc.) because people love data, especially about themselves (height, weight, heartbeat, blood pressure, etc.). The data is usually tracked manually but this is slowly becoming a thing of the past. With the development of our system, we are solving the problem of tracking and saving data of past sports activities. Among the existing solutions that we analyzed, we did not find a similar application concept. We discovered only functional journals to which we later added simplified or partially automatised data tracking. Consequently, we achieved so that our users can concentrate more on the sports activity itself, rather than remembering all of the data. This thesis presents the development and the implementation of an application for exercise recognition during a sports activity, as well as counting repetitions of activities on an Arduino Nano 33 BLE microcontroller. Furthermore, it presents the development of a mobile application developed in the Flutter framework, which enables a connection with said microcontroller through BLE connection and displays sent data. We have integrated our system with the Strava platform for data storage, which is collected and stored with the help of API requests. The result is a transparent mobile application with a simple and intuitive user flow and interface. For the application, we have built a basic model of machine learning that can discern a specific exercise activity and state of resting from movement, while also integrating a simple algorithm for counting repetitions. The system successfully simplifies the activity journal and therefore has room for improvement and growth.
|
