- 01Synthetic trace
- 02Gentle movement
- 03Count comparison
- 04Known limits
Pick a gesture you can reproduce
Use a gentle motion with a clear rest interval. Count attempts manually and compare them with accepted events. Record false counts and missed counts. A useful demo reports both instead of presenting only successful attempts.
The CoreS3 already includes the IMU and display needed for this project. If you developed the desktop policy first and want to build the physical version, the controller below is the relevant equipment. Extra sensor Units are not required.
Change one policy choice
Try a different entry threshold in the desktop tests, then deploy and compare a new physical trial. Keep the rearm threshold below the entry threshold. Record the chosen units and the library/board versions with your results.
This simple magnitude policy cannot distinguish a particular kind of gesture reliably across all users, mounting positions or motions. It also does not count steps accurately merely because a hand movement can trigger it.
Try it
Hold the signal above threshold in a synthetic replay and verify that waiting one second does not create a second event. Then return below the rearm threshold and make a new burst after the cooldown. This checks the difference between a timeout and an actual new edge.