Joint coordinate optimization in fingerprint-based indoor positioning

Abstract

Fingerprint-based indoor positioning uses pattern recognition algorithms (PRAs) to estimate the users' locations in wireless local area network environments, where satellite-based positioning methods cannot work properly. Traditionally, the training phase of PRA is separately conducted for x and y coordinates. However, the received signal strength from access points is a unique fingerprint for each measured point, not for x and y coordinates separately. In this letter, we propose a method to jointly employ the x and y coordinates during the training phase using a novel PRA-based Gaussian process regression (GPR), named 2D-GPR. Experimental results show that the proposed 2D-GPR improves the accuracy of positioning more than 40cm in limited data samples and has a lower calculation cost compared with conventional GPR.