User, device and orientation independent human activity recognition on mobile phones

Smart phones equipped with a rich set of sensors are explored as alternative platforms for human activity recognition in the ubiquitous computing domain. However, there exist challenges that should be tackled before the successful acceptance of such systems by the masses. In this paper, we particularly focus on the challenges arising from the differences in user behavior and in the hardware. To investigate the impact of these factors on the recognition accuracy, we performed tests with 20 different users focusing on the recognition of basic locomotion activities using the accelerometer, gyroscope and magnetic field sensors. We investigated the effect of feature types, to represent the raw data, and the use of linear acceleration for user, device and orientation-independent activity recognition.