This digital paddle improves desk tennis coaching

Desk tennis, AKA ping pong, is a severe sport and like all severe sports activities within the trendy age, its athletes can profit from data-driven coaching. With no coach watching, a desk tennis participant possible gained’t even have the ability to preserve observe of the variety of strokes made throughout a coaching session. To make that coaching extra productive, Samuel Alexander constructed this desk tennis digital paddle that analyzes strokes.

Alexander put severe effort into making the sensible paddle helpful for actual gamers and actual coaching, so it has correct steadiness and rebound. However embedded inside the deal with’s grip is an Arduino Nano 33 BLE Sense Rev2 board that displays the paddle’s motion. It makes use of that data to detect totally different sorts of strokes, retaining a tally of every that the participant can then evaluate on the OLED display screen. 

The paddle tracks motion by means of the Arduino’s built-in LSM9DS1 IMU. However the knowledge supplied by that sensor is simply an array of vectors and none of them correlates on to a stroke. To show the uncooked knowledge into stroke tallies, Alexander turned to Edge Impulse to coach a machine studying mannequin based mostly on real-world examples of every stroke sort.

It will probably now reliably classify regular movement (not a stroke), backhand drives, backhand smashes, forehand drives, forehand smashes, and forehand loops. The rely exhibits up on the grip’s OLED display screen, however Alexander additionally created an internet interface that customers can go to with their smartphones to see extra particulars, such because the session time and timestamps of strokes. 

Alexander discovered that this performs effectively, with an final take a look at accuracy of 88.7%.