Improved smartphone-based indoor localization via drift estimation for accelerometer

Indoor localization is a very promising application for smartphone. Normally, we can obtain the relatively accurate location information from GPS in outdoor circumstances. However, GPS will be invalid for indoor applications, and we cannot achieve indoor localization with raw data of inertial measurement units (IMU) due to the drift of accelerometer in smartphone. In this paper, we proposed an estimation algorithm based on Kalman filtering to estimate the constant drift of accelerometer with the measurement data obtained outside. Then we can get the relatively accurate acceleration indoors due to removal of the drift and measurement noise. Afterwards, the position would be obtained by the numerical integration of acceleration. Finally, the verification of the proposed algorithm was given via simulation with favorable estimated drift and accurate location information.