Efficient calibration for robust indoor localization based on low-cost BLE sensors

Abstract

The fingerprinting method for indoor localization is associated with a high calibration effort. In order to improve the efficiency of this time-consuming method, a new calibration procedure is proposed. It allows to reduce the calibration effort associated with fingerprinting considerably and ensures a sufficient accuracy at the same time. The proposed procedure takes into account the geometry of the environment subject to study, in order to empirically estimate the distribution of the signal strength at predefined positions. Thus, mobile sensor localization can be carried out robustly and the deviation from real position is less than 0.75 m in about 90% of the cases. This is achieved even though the calibration effort is reduced by almost 84% compared to the original one.