查看文章

We study the problem of recognizing the motion mode of users using solely signals from phone embedded GPS receiver and 3-D accelerometers. Motion mode recognition is motivated both by its usage in low-cost inertial aiding systems and as a preprocessing step for other motion mode dependent algorithms. Drawbacks of traditional inertial navigation can be mitigated by context information, enabling automatic zero-velocity updates and pedestrian dead-reckoning mechanization. As an example of other applications, we present 3-D acceleration based road condition monitoring system. The algorithms developed for this study are tested with real data collected with mobile phones, resulting in a realistic identification of limitations and possibilities of context-aware positioning applications.