Performance improvement of offline phase for indoor positioning systems using Asus Xtion and smartphone sensors

Abstract

Providing a customer with tailored location-based services (LBSs) is a fundamental problem. For location-estimation techniques with radio-based measurements, LBS applications are widely available for mobile devices (MDs), such as smartphones, enabling users to run multi-task applications. LBS information not only enables obtaining the current location of an MD but also pro- vides real-time push-pull communication service. For indoor environments, localization technologies based on radio frequency (RF) pattern-matching approaches are accurate and commonly used. However, to survey radio information for pattern-matching approaches, a considerable amount of time and work is spent in indoor environments. Consequently, in order to reduce the system-deployment cost and computing complexity, this article proposes an indoor positioning approach, which involves using Asus Xtion to facilitate capturing RF signals during an offline site survey. The depth information obtained using Asus Xtion is utilized to estimate the locations and predict the received signal strength (RF information) at uncertain locations. The proposed approach effectively reduces not only the time and work costs but also the computing complexity involved in determining the orientation and RF during the online positioning phase by estimating the user�s location by using a smartphone. The experimental results demonstrated that more than 78% of time was saved, and the number of samples acquired using the proposed method during the offline phase was twice as much as that acquired using the conventional method. For the on- line phase, the location estimates have error distances of less than 2.67 m. Therefore, the proposed approach is beneficial for use in various LBS applications.