To determine the outcome of the coin toss, we used Arduino Nano BLE 33 to analyze the result of coin tosses by using its built-in microphone to distinguish the two different sounds of the coin falling onto either the left or right side of the box. We trained a machine learning model in Edge-Impulse, using two series of sound data corresponding to the coin falling onto the two sides of the boxes and a series of sound data of noises, as shown below. This approach ensured that the CoinMind machine accurately predicted the outcome of each coin toss.