This week, we focused on enhancing both the functionality and polish of the adaptive music bike app. Here’s what we accomplished:
ML
- Introduced a file list viewer in the Sensors tab showing all saved
.txtfiles, complete with pull-to-refresh and delete options. - IMU data is now saved to timestamped .txt files in internal storage for clean, organized machine learning data collection.
- A new timeout mechanism that resets the last event after a period of inactivity—useful for cleaner signal processing.
FMOD
- Added Hall Effect direction (Forward/Reverse) to the UI and mapped it as an FMOD parameter.
- Enabled support for switching between multiple FMOD bank files, giving users greater control over music profiles.
Hardware/App
- PCB has been completed and ordered.
- App has a zero acceleromter button which replaces the physical zero button that existed on the hardware before
Once the PCB arrives we can finally begin testing on the bike and train the ML model.

Leave a comment