Humaine: a ubiquitous smartphone-based user heading estimation for mobile computing systems

Recently, there have been a wide range of mobile computing and crowd-sourcing applications that leverage the proliferating sensing capabilities of smartphones. Many of these place a paramount importance on accurate user heading estimation. Such applications include dead-reckoning-based localization and many crowd-sensing applications where the user typically carry her phone in arbitrary positions and orientations relative to her body and her transportation mode. However, there is no general solution available to estimate the user’s heading as current state-of-the-art focus on improving the phone orientation estimation, require the phone to be placed in a particular position, require a fixed transportation mode, require user intervention, and/or do not work accurately indoors. In this paper we present Humaine, a novel ubiquitous system that reliably and accurately estimates the user orientation relative to the Earth coordinate system. Humaine works accurately whether the user is riding a vehicle or walking indoors/outdoors for arbitrary cell phone positions and orientations relative to the user body. Moreover, it requires no prior-configuration nor user intervention. The system intelligently fuses the different inertial sensors widely available in off-the-shelf smartphones and employs statistical analysis techniques to their measurements to estimate the user orientation. Implementation of the system on different Android devices with 300 experiments performed at different indoor and outdoor testbeds shows that Humaine significantly outperforms the state-of-the-art in diverse scenarios, achieving a median accuracy of 14∘ and 16∘ for indoor and outdoor pedestrian users and 20∘ for in-vehicle users over a wide variety of phone positions. This is better than the-state-of-the-art by 523% and 594% for indoor and outdoor pedestrian users and 750% for in-vehicle users. This accuracy highlights the ubiquity of Humaine and its robustness against the various noise sources.