Implementation of a Somatosensory Interactive Game with Sensor and Neural Network Integration

This paper presents an interactive motion-sensing game designed for ocean protection, aligned with the United Nations' 14th Sustainable Development Goal. The game uses wearable gloves with accelerometer and gyroscope sensors for hand posture recognition, processed via sensor fusion and a feedforward neural network (FNN). Bluetooth Low Energy (BLE) with a 4-to-1 Central and Peripheral architecture facilitates communication, while the neural network function runs on the BLE microcontroller. The system development includes hardware, algorithms, and Unity game software implementation.