A Usability-Enhanced Smartphone Indoor Positioning Solution Using Compressive Sensing

Ubiquitous positioning is in highly demand in the fields of smartphone location-based services and more widely, in internet of things applications. Many WiFi fingerprinting algorithms have received widespread attention in the industry. Such methods for commercial location-based services however, still experience usability problems. These include low efficiency and high time costs when surveying radio maps. At the same time, heterogeneous radio measurements from different mobile devices may degrade positioning performance. This paper proposes a novel fingerprinting indoor positioning method aiming at breaking through the usability bottleneck in two ways: 1) a dynamic on-the-move surveying method of radio map to enhance the surveying efficiency using a self-developed mobile geo-reference system, and radio map is constructed using much sparser measurements and sparse principle. 2) Correspondingly, a sparse attribute based positioning algorithm was developed. This new positioning algorithm takes account of the heterogeneity of devices, and does not require any manual operations for the calibration of hardware bias, unlike many widely used static fingerprinting methods. The proposed positioning method was implemented on smartphones and verified in typical office indoor environments. The experimental results show that the radio map surveying efficiency was enhanced by 10 times in terms of radio map surveying time. Compared with the widely used static methods, the proposed positioning algorithm has more consistent positioning accuracy even when different devices were used for radio map surveying and positioning. The proposed fingerprinting positioning method is applicable when developing LBS applications (demo attached), while the enhanced usability will improve user experience of positioning performance.