Advanced Microwave Systems for Diverse Sensing Application

Abstract

As the Internet of Things (IoT) continues to expand across various industries, the ubiquity of connected devices is driving the widespread deployment of antennas in diverse indoor and outdoor environments. Among these, antenna sensors, a promising subset of microwave sensors, have gained significant attention for their unique ability to integrate both sensing and communication functions in a single device. As key components in the growing landscape of IoT and smart cities, sensor nodes require capabilities for both wireless communication and environmental monitoring. Antenna sensors offer a compact and cost-effective solution by combining these functions. Their advantages are manifold: integrating sensing and communication into a single antenna simplifies sensor node design by eliminating the need for additional sensor hardware. In addition, the commonly used microstrip antenna sensors can be manufactured using low-cost PCB milling machines. Their planar structure also facilitates easy integration with other RF components in the radio front end. Operating at microwave frequencies not only reduces the footprint of these sensors but also accelerates the absorption and desorption of target analytes, as microwaves can serve as an energy source. In summary, antenna sensors provide a robust, low-profile, and cost-effective solution for real-time wireless monitoring systems with communication capabilities. Their advantages are maximized when deployed in ubiquitous sensor networks and base stations, where many traditional sensors and antennas can potentially be replaced with antenna sensors. Given the wide range of applications for antenna sensors, they can be used either alone or functionalized with sensing materials in various scenarios, such as medical detection, food and crop classification, and volatile organic compound (VOC) sensing. With this motivation and background, this Ph.D. work on advanced microwave systems for diverse sensing applications is divided into three parts: The first part is a comprehensive summary of different types of radio frequency (RF)/microwave sensors based on a literature review of recent advances. The second part explores the implementation of cost-effective inkjet-printed flexible antenna sensors. The third part focuses on the development of compact antenna sensors based on nanomaterials and machine learning for various applications, including hand gesture recognition, VOC gas sensing, and liquid characterization.

The first part summarizes different types of RF/microwave sensors based on a literature review of recent advances. It explores five key application areas: food-related applications, agricultural applications, gas sensing, structural health monitoring, and biomedical/healthcare applications. As different applications require different types of antenna/microwave sensors, this section also delves into the operating principles of wireless microwave sensing. Finally, by analyzing the current limitations of microwave sensors, future perspectives are discussed, highlighting the potential of integrating advanced materials and machine learning to address complex real-world applications.

The second part focuses on the development of flexible antennas for ubiquitous conformal sensing applications, driven by the need for sensors that can be seamlessly integrated into a variety of non-planar surfaces and challenging environments. Such flexibility is essential for applications where traditional rigid antennas are impractical, including wearable technology, flexible electronics, and other conformal devices. As a proof of concept, inkjet printing technology is explored to create a slot dipole antenna. A compact dual-band flexible coplanar waveguide-fed stepped-impedance slot dipole antenna for ISM applications is developed. Using quarter-guided-wavelength stepped impedance resonator slots, this design achieves improved performance over conventional approaches. Experimental results demonstrate wide bandwidths in both the 2.4 and 5 GHz ISM bands, with stable performance under bending. Based on this promising result, a flexible antenna sensor for liquid detection is further developed. This novel sensor, designed for acetone/water detection, combines octagonal and square-shaped patches with a coplanar waveguide feed, inkjet-printed using nanosilver on Kapton polyimide. Operating in the 4.4-4.8 GHz range, the sensor exhibits high sensitivity to changes in acetone concentration and maintains functionality under deformation, making it suitable for conformal applications. The antenna also has good radiation characteristics, with a measured peak gain of 4.12 dBi and 76.5% efficiency, offering a compact, lightweight solution for applications such as cosmetics, industrial process control, and safety detection.

The third part addresses the challenge of eliminating the bulky and costly sensor arrays used in traditional sensing systems. Taking gas classification as an example, traditional e-nose systems consist of sensor arrays where each sensor is coated with different nanomaterials to create specific patterns or profiles for each gas under test. To streamline this design and reduce costs, a novel virtual sensor array (VSA) has been developed based on a single antenna sensor coated with a graphene oxide/Nafion composite for VOC detection. The sensor leverages multi-resonance properties to generate multi-dimensional data that serve as distinct fingerprints for multiple VOCs. Machine learning algorithms are employed to achieve exceptional accuracy in identifying multiple VOCs, including isomers, with a cross-validation accuracy of 91.7%. The sensor also accurately determines concentrations in binary VOC isomer mixtures with an R-squared value exceeding 0.98. Notably, the antenna maintains stable communication performance throughout the sensing operation, overcoming key challenges in antenna-based VOC sensing. This innovative approach offers a promising solution for dual-functional sensor nodes in various VOC-related applications within IoT networks. Furthermore, building on the antenna sensor developed in this part, the multifunctionality of the sensor is explored using deep learning in the fourth part. Motivated by the goal of making the antenna sensor capable of both environmental sensing and human-machine interaction, the integration of multiple functions into a single piece of antenna significantly streamlines the system. The proposed multifunctional antenna sensor integrates hand gesture recognition, VOC gas identification, permittivity characterization, and fixed-band communication on a single 33×33 mm substrate. Utilizing ultra-wideband feature extraction and efficient machine learning algorithms, the sensor achieves 94.1% accuracy in recognizing ten hand gestures and 100% accuracy in identifying six VOC gases, including isomers. The sensor also successfully characterizes different liquids and substrate blocks without affecting the communication band. This innovative approach addresses key challenges in IoT sensor nodes, offering a compact, cost-effective solution that combines sensing and communication capabilities.

The results of this Ph.D. work will pave the way for cost-effective compact antenna sensing systems that will play an important role in a variety of applications, including environmental sensing and human-machine interaction.