Detection of critical activity by accelerometer and gyroscope sensors

In this study, a hardware prototype that has an inertial measurement unit including an accelerometer and a gyroscope with the ability to record sensor readings is developed. Using the prototype on the training weapons, gyroscope and accelerometer readings were recorded, during gunshots as well as other movements namely, shocks, shaking and random movements. Later, a method that detects the gunshots and classifies them against shocks, shaking and random movements is proposed. The proposed method is based on extracting features from signals generated by the different movements, and identifies them using a classifier. The spectral sub-band energy ratios represent the movements and the k-Nearest Neighbor algorithm is used for classification. It is shown that, gunshot movements are recognized correctly against shocks, shaking and other random movements with a true classification rate of 98%. In addition, this study can also provide input to analyzes of characteristic movements of human and mechanical systems.