Auto-calibration for device-diversity problem in an indoor localization system

Recently, the techniques for indoor localization have become more and more important and play a critical role in many mobile applications. Among them, the fingerprint-based indoor localization system has been recognized as a possible right way toward success. However, some challenges still remain. One issue should be addressed is the device diversity problem, where different devices would receive different radio signal strengths (RSS) at the same location. This problem breaks the fingerprint assumption - each location has its singular RSS. Conventional calibration methods require manually collecting pair-wise RSS data among devices to train the calibration model. To reduce human load, we proposed a method that could automatically calibrate the device diversity problem in an efficient way.