TaLc: Time Adaptive Indoor Localization with Little Cost.

Liqiong Chang,Xiaojiang Chen,Ju Wang,Dingyi Fang,Chen Liu,Zhanyong Tang,Weike Nie

DOI: https://doi.org/10.1145/2799371.2799380

2015-01-01

Abstract:ABSTRACTMany emerging applications and the ubiquitous wireless signals have accelerated the development of WiFi-based indoor localization techniques. However, there is one primary issue, that is the environmental dynamics will greatly undermine the fingerprint stability over time, hindering these techniques to be fully practical. Although many researchers dedicated to deal with the fingerprint adaptation problem, it still remains an open problem without a wholesome approach. In this study, we propose TaLc, a time adaptive and low cost indoor localization approach that fortifies the indoor localization to be a fully practical service. TaLc can automatically identify the time to update the fingerprint without the need of any extra devices. Then by collecting the fresh RSS measurements of only a few of reference locations, which are at most equal to the number of APs and much less than the number of locations, to update the fingerprint and thus reduce the human effort for fingerprint calibrating. Finally TaLc utilizes the Compressive sensing (CS) based localization to reduce the data volume and energy consumption, which can achieve an fine-grained localization accuracy with only a few of RSS measurements. Extensive real-word experiments over 3 months illustrate that TaLc can effectively update the fingerprint with the average error of less than 3.6dBm and improves the localization accuracy of about 50%.