A Calibration-Free Crowdsourcing-Based Indoor Localization Solution

Researches on crowdsourcing-based localization systems have been attracting much attention. It is a main problem that device diversity and short-duration signal strength measurement significantly degrade the localization accuracy in crowdsourcing-based systems. In this paper, we analyze underlying relationships between detected wireless Access Points (AP) and received signal strength (RSS), which are relatively invariable over devices and measurement times. Then we present a novel solution which uses these underlying relationships as key values for location determination. We use the first publicly available database in this field to evaluate this solution. The experimental results confirm that this solution provides high success rate and acceptable localization accuracy.