AAU Student Projects - visit Aalborg University's student projects portal
A master's thesis from Aalborg University
Book cover


Sensing the Orientation of Passive UHF Tags Using a Commercial RFID System

Authors

; ;

Term

10. term

Publication year

2011

Submitted on

Pages

104

Abstract

Denne afhandling undersøger, om objekters orientering kan estimeres ved hjælp af almindelige, passive UHF RFID-tags i et kommercielt system, så man kan opdage fx om en kasse vender “rigtigt” i en forsyningskæde uden at tilføje dyre sensorer. To metoder udvikles: en bayesiansk Probabilistic Orientation Algorithm og en ikke‑parametrisk KNN Orientation Algorithm. Begge omfatter en kalibreringsfase, hvor systemet lærer forventede modtagne effektværdier, og en estimeringsfase, hvor orienteringen udledes fra de målte signaler. Metoderne valideres først med et kontrolleret setup (kunstigt tag og spektrumanalysator) og evalueres derefter med en kommerciel RFID‑læser og passive tags; resultaterne viser høj nøjagtighed ved estimering af objekters hældning og demonstrerer gennemførlighed med standardudstyr. Arbejdet behandler desuden kanal‑ og polarisationsforhold, præsenterer tiltag til at reducere omkostninger (kortere kalibrering og færre antenner) og undersøger online, adaptiv kalibrering med referencetags for at kunne fungere i dynamiske miljøer.

This thesis investigates whether the orientation of objects can be estimated using standard passive UHF RFID tags within a commercial system, enabling detection of conditions such as “this side up” in supply chains without adding costly sensors. Two methods are developed: a Bayesian Probabilistic Orientation Algorithm and a nonparametric K‑Nearest Neighbor (KNN) Orientation Algorithm. Both include a calibration phase, in which the system learns expected received power values, and an estimation phase that infers orientation from measured signals. The methods are first validated in a controlled setup (artificial tag and spectrum analyzer) and then evaluated with a commercial RFID reader and passive tags; results indicate high accuracy in estimating object inclination and demonstrate feasibility with off‑the‑shelf equipment. The work also addresses channel and polarization effects, presents approaches to reduce costs (shorter calibration and fewer antennas), and explores online adaptive calibration with reference tags to support operation in dynamic environments.

[This summary has been generated with the help of AI directly from the project (PDF)]