We created a minimum viable product (MVP) that is capable of collecting data and classifying it.
The hardware consists of the following parts:
- Peripheral system
- 3d-printed box
- Arduino Nano BLE 33 Sense board
- Perf board
- 9V battery and battery holder
- Switch
- Button
- Watch band
- Central system
- 3 Arduino Nano BLE 33 Sense boards
- Laptop computer
- 3 USB cables
A minimum viable product (MVP) was created. It is split up into two systems, the peripheral system and the central system. The peripheral system is the part that is attached to the user’s arm as he punches. The central system is connected to a laptop computer and does the machine learning required to classify the punches. The peripheral system is composed of a 3D-printed box, one Arduino Nano board, a perf board, a 9V battery and battery holder, a switch, a button, and a watch band. The 3D-printed box is made of PLA and serves as a holder for the rest of the components in the peripheral system. It has a sliding lid that protects the components inside from getting destroyed. The Arduino Nano board in the peripheral system collects the punching data and sends it via Bluetooth to the central system. The peripheral system is powered by the 9V battery, which is held by the battery holder. The batter holder is important, because it allows the battery to be replaced when it dies. The switch turns the system on or off by controlling the power sent to the Arduino Nano, and the button resets the board. The perf board is used to connect all of the electrical components. The battery, switch, and button are all connected to perf board using jumper cables that were soldered to each part. The watch band attaches the box to the user’s forearm.



The central system is composed of three Arduino Nano boards, a laptop computer, and three USB cables. One of the Arduino Nano boards receives the data from the peripheral board via Bluetooth. The other two Nanos utilize TinyML to classify the data – one runs the binary classification model and the other runs the punch classification model. All three of the central Arduino Nanos are connected to the laptop computer using the USB cables. The final block diagram for the hardware system is shown below.
