Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Intelligent wireless networks and robots for low-cost battery-less sensing and localization

Giannelos Evangelos

Full record

---

Summary

This dissertation presents a comprehensive exploration of radio frequency identification (RFID) technology merged with robotics, shedding light on their transformative potential and inherent challenges. At the heart of our research lies the meticulous study of monostatic localization using novel techniques that do not necessarily imitate a synthetic aperture radar (SAR), in sharp contrast to prior art; instead, the adopted approach exploits sparse and low-bandwidth phase measurements with appropriate particle filtering weights, so that robustness to indoor multipath is achieved. As a result, state-of-the-art performance in 3D localization utilizing commercial Gen2 RFID tags is achieved with a mean absolute error of 24 cm.Building upon this foundational work, the research delves into the intricacies of multipath effects, particularly the distortions introduced by environmental reflections. Notably, the investigation in this realm resulted in a technique that jointly estimates the location of the tag as well as its reflector with a reflector localization accuracy of 5 cm; such (perhaps) pioneering endeavor adheres to low bandwidth limitations of existing, commercial RFID systems.Furthermore, study of bistatic/multistatic localization is performed, offering elliptical direction-of-arrival (DoA) estimation and 2D/3D localization techniques, with unique geometrical considerations and their implications on RFID systems. These techniques offer estimations with as low as 5° and 9 cm mean absolute error for DoA and 2D localization respectively.Transitioning next into the realm of wireless sensor networks (WSNs), the dissertation elucidates the symbiotic relationship between RFIDs and WSNs, emphasizing key considerations surrounding interrogation architectures and their real-world ramifications. Experiments include both unmanned aerial vehicles (UAV), as well as ground-based pedestrian interrogation systems, using bistatic and monostatic architectures.A paramount feature of this research is the seamless integration of theoretical studies with tangible applications, resulting in real-world systems primed for diverse applications. This is exemplified in our development of a mobile interrogation and localization system, integrated with the Robot Operating System (ROS), capable of pinpointing RFID tags in complex environments, like warehouses, libraries, and offices. Further, the synergy of UAV-based interrogation with backscatter technology emerges as a beacon of innovation, signaling new horizons in precision agriculture.

Available Files

Services