Conventional nondestructive testing techniques do not suffice to efficiently inspect the defects in multiple nonmetallic pipes. Thus, in this article, we propose a novel method for inspection of these components based on near-field microwave holographic imaging augmented with beamspace transformation. This approach harnesses the ability of microwaves for 3-D imaging by measuring the backscattered fields in the near-field region. In the proposed technique, we augment the solution of near-field holographic imaging through the use of beamspace transformation which allows for focusing on each pipe during the solution process. This improves the quality of the reconstructed images, in particular, for the pipes which are further away from the measuring antennas. First, we demonstrate the performance of the proposed imaging technique using simulations. Then, we present the experimental results for imaging of concentric cylindrical polyvinyl chloride (PVC) pipes using wideband microwave data.

Detection of pollutants in water is conventionally performed in wet chemistry labs for samples that are collected manually. This task is time-consuming, costly, and difficult to expand to large water bodies and perform frequently. Thus, to facilitate water quality testing, a new methodology is presented based on the use of a non-uniform array of microwave sensors and applying machine learning to the collected data. The sensor elements resonate at different frequencies which cover a broad bandwidth, providing sufficient information for machine learning algorithms to determine the type of pollutant. Here, as a proof-of-concept, the proposed methodology is tested with water samples including Phosphate (PO ₄ ), Lead (Pb), Mercury (Hg), and Chromium ⁺⁶ (Cr ⁺⁶ ) with various concentrations. The proposed technology is fast, cost-effective, repeatable, and expandable for detecting a larger number of pollutants.

Gesture recognition, in particular, hand gesture recognition, has received great attention in the field of human-computer interaction. Although several hand gesture recognition designs have been proposed in the past, they still suffer many limitations in practice. Hence, we present a wearable device, called GloveSense, which is made on a regular glove using conductive threads. The system consists of ten coils, with each pair of coils sewn on both sides of each finger and connected in series. To achieve high sensitivity, a data acquisition system is employed in which a tank circuit made of each sewn coil is measured. Since the utilized commercial data acquisition board has limited channels, a switching mechanism is employed to measure some of the extra coils. The utilized switching mechanism reduces the cross-talk effect between the readings. The GloveSense system is tested for recognition of numbers 1 to 10 in American Sign Language. A dataset of 1400 responses is employed along with a machine learning algorithm for gesture recognition. The proposed system demonstrates a high accuracy of 99% and a short recognition time of less than 1 second.

Schizophrenia is a serious mental illness associated with neurobiological deficits. Even though the brain activities during tasks (i.e., P300 activities) are considered as biomarkers to diagnose schizophrenia, brain activities at rest have the potential to show an inherent dysfunctionality in schizophrenia and can be used to understand the cognitive deficits in these patients. In this study, we developed a machine learning algorithm (MLA) based on eyes closed resting-state electroencephalogram (EEG) datasets, which record the neural activity in the absence of any tasks or external stimuli given to the subjects, aiming to distinguish schizophrenic patients (SCZs) from healthy controls (HCs). The MLA has two steps. In the first step, symbolic transfer entropy (STE), which is a measure of effective connectivity, is applied to resting-state EEG data. In the second step, the MLA uses the STE matrix to find a set of features that can successfully discriminate SCZ from HC. From the results, we found that the MLA could achieve a total accuracy of 96.92%, with a sensitivity of 95%, a specificity of 98.57%, precision of 98.33%, F1-score of 0.97, and Matthews correlation coefficient (MCC) of 0.94 using only 10 out of 1900 STE features, which implies that the STE matrix extracted from resting-state EEG data may be a promising tool for the clinical diagnosis of schizophrenia.

In this paper, a novel methodology is proposed for material identification. It is based on the use of a microwave sensor array with the elements of the array resonating at various frequencies within a wide range and applying machine learning algorithms on the collected data. Unlike the previous microwave sensing systems which are mainly based on a single resonating sensor, the proposed methodology allows for material characterization over a wide frequency range which, in turn, improves the accuracy of the material identification procedure. The performance of the proposed methodology is tested via the use of easily available materials such as woods, cardboards, and plastics. However, the proposed methodology can be extended to other applications such as industrial liquid identification and composite material identification, among others.