An Indoor Positioning System Using Vision aided Advanced PDR technology without image DB and with motion recognition

Abstract

In this paper, we proposed an indoor positioning system using vision aided advanced PDR technology. The novelty of this research is that the vision-based positioning system does not need an image DB and the PDR system could estimate a position regardless of any motions. For motion recognition, we make an ANN-based classifier and which gives a very high recognition rate. We estimate a height using barometric pressure sensor calibrated by sea-level pressure and true altitude. We experiment a field test and its result is less than 3% of error rate. We anticipate that the proposed system will be applied to the smartphone-based navigation system.