|Abstract:||With the growing populations of smartphone users in daily life, the navigation based on the mobile device is rapidly developing in both academic and industrial fields. When it comes to navigating with handheld mobile devices, additional information is commonly utilized for increasing the accuracy. In order not to be constrained by sensor integrations of those outside measurements, navigation with built-in inertial sensors only is a matter of interest in the paper. Finding a user’s location with angular rate update, however, diverges by time, so accelerometer and magnetometer are used for error compensation. Especially in case of magnetic disturbance, the sensor output is easily affected by outside environment, so the algorithm regarding the perturbation is necessary. In this paper, the magnetic anomaly detection algorithm for estimating smartphone attitude is proposed based on a machine learning method. The magnetically disturbing condition is determined after the device motion classification beforehand using the learning technique. Whether the smartphone is stationary or moving is determined, the magnetic anomaly of current time step is classified in turn through k-Nearest Neighbor(kNN) or Support Vector Machine(SVM). The classification steps in both action and magnetic field disturbance are described in the paper, and the simulation and experimental results show that the proposed algorithm is able to determine the magnetic distortion area.|
Proceedings of the 29th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2016)
September 12 - 16, 2016
Oregon Convention Center
|Pages:||1184 - 1189|
|Cite this article:||
Park, So Young, Heo, Se Jong, Park, Chan Gook, "Detection of Magnetic Anomaly Based on a Classifier for Smartphone Attitude Estimation," Proceedings of the 29th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2016), Portland, Oregon, September 2016, pp. 1184-1189.
ION Members/Non-Members: 1 Download Credit