An RFID-Based Object Localization Framework and System

Abstract

Locating objects is a key requirement in several of the emerging computing paradigms. The problem of locating objects has been extensively studied from a variety of technological and technique-oriented perspectives. Recently, Radio Frequency Identification (RFID), a wireless automated identification technology, has come forth as a viable platform for locating objects, particularly in indoor environments. While rapid advances in RFID-based object localization are evident, current approaches lack adaptability, reliability, and scalability. This thesis addresses these issues and presents an RFID-based object localization framework and system to help locate stationary and mobile objects with high accuracy. Our RFID-based object localization framework and system is resilient in select environmental conditions, accommodates numerous use-case scenarios, and is tag orientation and vendor hardware –agnostic. We demonstrate that radio signal strength, a technique used in our location system and traditionally considered unreliable, can be used as a reliable metric for locating objects in selective cases. Additionally, we show that tag sensitivity caused by manufacturing variation influences object localization performance and we present tag selection and binning techniques. This ensure range and cost -optimized uniformly sensitive tags, leading to a reliable and high-performance object localization. We further improve the object localization characteristics of our system by matching tags to readers and demonstrating that reference tags could be made optional without significant loss in performance. Rigorous experimental evidence suggests that our RFID-based object location system can simultaneously locate several stationary and mobile objects in realistic noisy indoor environments with localization accuracy in the range of 0.15-0.84 meters. We have also developed several visualization applications focusing on a variety of computing platforms to help visualize the targeted object’s location.