Evaluation of cross-assistive approach for PDR and Wi-Fi positioning in the real environment

Hybrid indoor positioning researches, which aim improving the positioning accuracy by applying multiple positioning techniques, are the focus of many researchers in these days. However, the current approaches do not consider the positioning accuracy, and tend to combine the positioning techniques together without further caring about the accuracy of each technique. We propose a novel method named "cross-assistive approach" for improving the positioning accuracy by using PDR and Wi-Fi positioning in a cross-assistive manner, where both the positioning techniques mutually increase the positioning accuracy by complementing each other's positioning result, thus we can estimate a user's real walking trajectory with high accuracy. In the cross-assistive approach, Wi-Fi positioning detects the initial values of PDR and corrects the accumulated error of PDR. Moreover, PDR improves Wi-Fi positioning as well. In our previous research, the initial parameters for PDR are manually configured. However, we solved the problem by automatically correcting the current position and direction using the estimated values from Wi-Fi positioning. We performed experiments in the real environment in the Umeda underground shopping mall in Osaka. As a result of the evaluation, we could estimate the initial values of PDR with the probability of 84% using Wi-Fi positioning, and improved the positioning accuracy up to about 10 meters.