A Practical Indoor Localization System with Distributed Data Acquisition for Large Exhibition Venues

Abstract

In this paper, we focus on the Wi-Fi based indoor localization in large exhibition venues. We identify and describe the real-world problems in this scenario and present our system. We adopt a passive way to detect mobile devices with the consideration of users’ preference and iOS devices’ privacy issue, and collect signal strength data in a distributed manner which meets the practical demand in exhibition venues and save the power consumption of mobile devices. Since exhibition venues have many restrictions on traditional localization approaches, we propose our approach and solution to fit these special conditions. We propose the clustering and Gaussian process regression (GPR) method to improve localization accuracy. Series of experiments in Hong Kong Convention and Exhibition Centre (HKCEC) show our system’s feasibility and effectiveness. Our approach has significant improvement in the localization accuracy when compared with traditional trilateration, fingerprinting and the state-of-the-art approaches.