Classification of physical activities using accelerometer signals

In this study, basic physical activities are classified using 3-axis accelerometer signals. The system is mainly proposed for mobile devices and basically objected to detect whether a person carrying this hardware is walking, running, or travelling in a car or bus. The developed system considers the accelerometer signals in constant bins and performs a real-time classification. The proposed system correctly classifies the signals obtained from a walker or from a person traveling in a vehicle at the rates of 95.66% and 99.23%, respectively.