A method of pedestrian dead reckoning for smartphones using frequency domain analysis on patterns of acceleration and angular velocity

In recent years, as smartphones have widely spread in the globe, it can be said that billions of users are carrying with high performance computer being always powered on. Since most of the advanced smartphones are equipped with motion sensors (accelerometers, gyroscopes and magnetometers) which are prerequisites for pedestrian dead reckoning, or PDR, it can be highly expected to realize seamless indoor/outdoor localization of the pedestrian users. However, the smartphones can be held in many ways such as in users' hands, pockets or handbags and they can be also subject to several complicated motions other than walking locomotion. In this research, we have aimed at providing technologies to estimate location and orientation (attitude) of the smartphones in users' hands, pockets and handbag even in the presence of several complex motions such as one caused by swinging his/her arms or the handbag. Since human walking locomotion has discriminative and repeating patterns in the components of acceleration and angular rate in the moving (walking) direction and the vertical/lateral direction even in the presence of other additional motions such as one caused by swinging arms, these patterns can be detected and extracted in the frequency domain analysis.