Walking Detection Using the Gyroscope of an Unconstrained Smartphone

In recent years, mobile devices (e.g., smartphones, tablets and etc.) equipped with various inertial sensors have been increasingly popular in daily life, and a large number of mobile applications have been developed based on such built-in inertial sensors. In particular, many of these applications, such as healthcare, navigation, and etc., rely on the knowledge of whether a user is walking or not, so that walking detection thus has attained much attention. This paper deals with walking detection by using the gyroscope of any commercial off-the-shelf (COTS) smartphone, which can be placed at different positions of the user. Inspired by the fact that the walking activity often results in notable features in the frequency domain, we propose a novel algorithm based on fast Fourier transformation (FFT) to identify the walking activity of a user who may perform various activities and may hold the smartphone in different manners. A thorough experiment involving three testers and multiple activities is carried out and confirms that the proposed algorithm is superior to the existing well-known counterparts.