Ultrawide band (UWB) techniques have become very well established for short range object detection particularly in applications such as ground penetrating radar for detection and location of buried targets, proximity monitoring and through-wall imaging in security, and monitoring vital signs of the human body for instance respiration and heart rate. The presentation will consider recent advances in UWB sensing in particular relating to (i) the use of UWB synthetic aperture radar techniques and (ii) UWB polarimetry.  Results will be shown to demonstrate the potential of these advances in applications such as high stand-off ground penetrating radar from mobile platforms, target characterisation, non-destructive testing and medical diagnostics.

This presentation introduces a new simulation framework for Frequency Modulated Continuous Wave (FMCW) radars, aimed at accurately modeling important radar characteristics like time-domain raw data, antenna arrays, multipath propagation, and radiation patterns, all within the Unity game engine. The framework uses a manual ray tracing method combined with Gaussian function-based modeling of the radar’s radiation pattern, allowing detailed control over key FMCW radar settings such as chirp time, slope, bandwidth, antenna configuration, gain, beamwidth, and side lobes. The simulation’s accuracy is tested in a complex parking garage filled with metallic structures that cause significant multipath effects. By closely replicating this challenging environment, the simulation demonstrates the benefits of accurately modeling radiation patterns and multipath reflections, leading to improved radar point cloud quality. The results show a strong alignment between the simulated data and real radar measurements, with clear improvements in point cloud accuracy due to the refined handling of multipath effects and radiation patterns. This validation underlines the framework’s potential as a useful tool for radar system design and research, particularly in difficult environments like parking garages.

Microwave sensing and heating technology has made considerable progress in the last 10 to 15 years and has developed into an established field of research with clearly defined opportunities and research directions. This field now encompasses a wide range of topics, including sensor development, heater optimisation, advanced manufacturing processes and the integration of these devices into complex systems. Researchers and engineers are actively exploring these opportunities to push the boundaries of microwave technology. In this presentation, we will look at the latest developments in microwave heating and sensing, with a particular focus on nanolitre and microlitre scale applications. The discussion will focus on devices developed using planar and waveguide technologies that offer unique advantages in terms of precision and efficiency. We will explore the intricate process of manufacturing these devices and highlight the challenges and innovations involved. In addition, the talk will cover the seamless integration of these microwave devices with fluidics technologies, highlighting their potential in various fields such as biomedical research, chemical analysis and materials science.

Soil electromagnetic parameters are related to physical and chemical properties, and mapping within-field variability provides valuable information for precision agriculture applications. Proximal soil sensing is convenient for this purpose due to its high spatial resolution. In the high-frequency electromagnetic induction modality for soil sensing in the frequency range from 3 MHz to 30 MHz, it is possible to measure both electrical conductivity and dielectric permittivity. The derived analytical model of the sensor above the double-layer homogeneous medium shows that the sensitivity to electrical conductivity and dielectric permittivity increases with frequency while the sensitivity to magnetic susceptibility is unchanged with frequency. The designed soil sensor consists of a shielded printed circuit board transmitter and receiver coils, which are separated and mutually perpendicular. The soil sensor was evaluated in the laboratory for the ranges of electrical conductivity (10 mS/m to 300 mS/m) and dielectric permittivity (3 to 40), typical for most soil types. The results agree with the numerical and analytical model and demonstrate the possibility of simultaneous electrical conductivity and dielectric permittivity measurement.

In this work, we demonstrate that the ripeness of avocado fruit can be analysed based on frequency-dependent electrical conductivity using a non-invasive and non-contact magnetic induction spectroscopy (MIS) method. Bioimpedance spectroscopy is useful electrical impedance property measurement of biological sample over a specific range of frequency, which can genrate information on physical, chemical or other cellular and molecular properties of the target sample. By utilizing an MIS sensor to perform individual body bioimpedance measurements on avocado fruit, scanning across multiple frequencies allows for the clear observation of dispersion behaviour due to capacitance effects at cell boundaries. For MIS, although non-contact measurement conveys some distinct advantages, selectivity remains a challenge. By selecting factors such as temperature, ripeness, pH value, firmness, ionic concentration, and other analysable and quantifiable attributes, we can establish quantifiable correlations between the bioimpedance spectrum and the ripeness of the samples and explain the reasons for bioimpedance changes during physiological process. Despite the variations among individual samples, batch testing can still demonstrate statistical significance for assessing ripeness.

In an everyday scenario, medical devices used for biomonitoring need to be energy efficient but also ensure the security of the private data they record. Capacitive intrabody communication is a method that enables secure and energy-efficient data transmission between devices on and inside the human body, but is susceptible to environmental changes. In this work, we use a liquid and gel phantom representing human tissue to measure and characterize the communication channel between an in-body and an out-body device of an implantable capacitive intrabody communication system.

The magnetic polarizability tensor (MPT) provides a means of quantifying the perturbation of a time-varying magnetic field due to the presence of a metallic object and is highly correlated with the object’s geometrical and material properties. As such, MPT is an essential tool for metallic object characterization using electromagnetic induction sensing. This presentation presents a novel approach to MPT measurement using a pulse induction metal detector in combination with an electromagnetic tracking system. Machine learning-based depth estimation is used to enhance the performance of the standard nonlinear least squares (NLS) inversion method. The proposed algorithm is experimentally validated in laboratory conditions. A significant improvement in measurement repeatability over the standard NLS inversion indicates the great potential of the proposed approach for hidden metallic object detection.

This talk will provide an overview of advanced electromagnetic sensing instruments and signal processing approaches for detecting and identifying buried explosive hazards. Subsurface explosive hazards fall into three main categories: 1, metallic (such as unexploded ordnance and metallic mines); 2. low-metal-content mines and mines constructed with intermediate conductive materials, and 3. plastic mines. Over the past two decades, significant strides have been made in developing advanced electromagnetic induction (EMI) devices and signal processing approaches. These EMI devices have been designed, built, demonstrated, and transferred to the commercial realm for the detection and identification of unexploded ordnance (UXO) along with advanced signal processing methods. Building on the success of detecting and classifying subsurface targets, research has shifted towards high-frequency EMI sensing for detecting and identifying low-metal-content targets, short and long command wires, intermediate electrical conducting targets, as well as plastic targets. In this talk, we will first describe the basics of advanced electromagnetic induction sensing technologies for geophysical applications. This will include recent advancements in EMI instruments and software developed to detect, locate, map, and identify subsurface munitions, along with demonstrations of blind classification results at live UXO sites. The classification results will be demonstrated for targets found in both magnetic soils and marine environments. Secondly, we will outline high-frequency electromagnetic induction sensors and signal processing approaches for detecting landmines, improvised explosive devices, non-metallic projectiles (such as smart bombs), and subsurface infrastructures (such as utility wires and pipes). Finally, the integration of electromagnetic sensing technologies on remotely controlled systems, such as unmanned ground vehicles (UGVs), unmanned aerial systems (UAS) drones, and autonomous underwater vehicles (AUVs), will be highlighted.

The accurate modelling of the eddy-current defect interaction has become an indispensable tool for the design and optimisation of eddy-current sensors. Despite the striking recent advances in modern generic numerical solvers, notably the finite element method, the application of such solvers for treating the eddy-current response problem is sub-optimal. In particular, the eddy-current response problem is characterised by a number of peculiarities such as small lift-offs and the presence of strong field gradients, which makes the use of mesh-based solvers quite delicate for a number of applications. Variational approaches, on the other hand, involving the integral equation formalism are better suited since they are directly concerned with the field perturbation in the defect support. Their application though requires the construction of dedicated integral kernels, which intrinsically take into account the nominal piece geometry. The construction of such kernels is, however, far from being trivial. In addition, although the integral equation formulation for the harmonic problem has reached a satisfactory maturity level, their application in transient calculations is certainly much less studied. In this talk, an overview of recent advances in variational methods for calculating the defect response will be presented, and important specialisations for the treatment of important flaw types (corrosion, pitting, stress-corrosion cracking) will be discussed. Important features such as the coupling with the measuring circuit will be evoked. Finally, different generalisation approaches for the pulsed eddy-current signature calculation will be presented.

The calculation of the magnetic field distribution produced by a coil in a ferrite U-shape core is an important factor in the analytical approaches for using such probes in Eddy Current Testing (ECT). The first step is to ensure the reliability of the simulation results. Hence, this study presents a comparison between measurement data and numerical simulation of the same geometry. A Hall effect sensor is used to measure the vertical magnetic field Bz along a U-shape high-permeability ferrite core, utilizing a Lock-in amplifier. For the simulation, the commercial FEM software COMSOL Multiphysics is employed. The very good agreement between experimental and numerical results constitutes the first step towards further research using a simplified geometry, which is more analytically describable, to minimize the requirements and difficulties of analytical approach models.

Ferrite-core probes are commonly deployed in eddy-current non-destructive evaluation problems, and various studies pursue the construction of deterministic models for such configurations. In this work, we develop a surrogate model for the problem of an eddy-current probe with ferrite core over a conductive specimen based on polynomial-chaos (PC) theory, which is suitable for performing uncertainty-quantification and sensitivity-analysis studies. Based on a number of deterministic finite-element simulations conducted with COMSOL, a polynomial surrogate can be constructed when geometric dimensions (e.g. the core’s vertical position, the probe’s lift-off, etc) and magnetic parameters (such as the core’s magnetic permeability) of the investigated problem are treated as random variables. Once the surrogate model is completed, statistical metrics and sensitivity indices of the quantities of interest (e.g. magnetic field) can be easily computed via post-processing. To reduce the computational cost, the sparsity of the model is promoted by employing sparse solvers for the extraction of the PC-expansion coefficients, while a sequential sampling technique based on variance criteria facilitates the selection of the most appropriate samples from the random space. The reliability of the computations is assessed against reference results obtained via the - computationally expensive - Monte Carlo method, and the computational savings compared to standard PC implementations are estimated.

In the context of Non-Destructive Evaluation (NDE) for Carbon Fiber Reinforced Polymer (CFRP) materials used in space applications, significant advancements necessitate the integration of Full Matrix Capture (FMC) for data acquisition, the Total Focusing Method (TFM) for signal analysis, and machine learning (ML) algorithms for enhanced defect recognition. This talk presents a study that explores the implementation and evaluation of these advanced techniques. FMC enables exhaustive data collection by utilizing every element in an ultrasonic array as both a transmitter and receiver. TFM processes this data to produce highly detailed images, allowing for precise defect characterization. The study incorporates anisotropic velocity profiles into the TFM algorithm, addressing the directional variations in ultrasonic wave velocities within CFRP materials, which significantly improves image quality. Moreover, the integration of machine learning algorithms, specifically diffusion maps, enhances the defect recognition process. These algorithms effectively handle noise and offer computational efficiency, making them well-suited for real-time applications. The study demonstrates the effectiveness of these combined techniques in detecting and characterizing artificial defects within CFRP specimens, highlighting their potential for improving the safety and reliability of space structures.

Frequency-domain features of ultrasonic acoustic emission (AE) signals are often used for fault detection or preventive maintenance in many long-term field monitoring NDT applications, such as crack, leak, or cavitation detection in vehicles, industrial machinery, aerospace, energy storage and transport, structural health monitoring etc. This requires continuous real-time processing, posing limitations to traditional AE bulk-PZT transducers and signal acquisition equipment in terms of size and power-consumption. Therefore, we explore a concept of a novel MEMS-based ultrasonic AE NDT contact sensor, implementing an in-sensor, "zero-power" signal frequency decomposition functionality. In the first part of the talk, we describe our preliminary research efforts on designing arrays of passive piezoelectric micromechanical ultrasonic frequency-selective cantilever resonators, for multi-channel band-pass signal filtering, in PiezoMUMPs SOI process. Next, we show an accompanying multi-channel low-power design of analog CMOS sensor interface for signal conditioning and detection. Finally, we show the design of digital circuits for logging, storage, and readout.

This presentation explores the challenges faced when utilizing autonomous underwater vehicles (AUVs) equipped with RGB cameras to automate the task of underwater marine debris detection. The underwater domain presents unique challenges due to the distortion and attenuation of light impacting the underwater image formation process and resulting in degraded image quality. An overview of the underwater image enhancement and image quality assessment methods will be given. Special focus will be put on how image quality affects the task of underwater marine debris detection, supported by insights from the Seaclear Marine Debris Dataset. Difficulties in evaluating model robustness and generalizing the performance of deep learning-based models to variable underwater conditions will be considered. Finally, we will briefly discuss innovative sensor combinations and options that could further enhance underwater marine debris detection.

Providing accurate information about all objects in the environment surrounding an autonomous vehicle is essential for the development of a maritime navigation system used on board autonomous and unmanned vessels. In this work, we introduce a method that integrates LiDAR and camera data for precise 3D object detection and tracking. The approach utilizes LiDAR for acquiring object position information and ensures accurate object classification through data from the camera. The proposed architecture will be presented with a detailed explanation of all algorithms used. Several scenarios have been created in a simulated environment to compare three approaches for the selection of target object in the camera frustum (the Nearest Center, Gaussian Likelihood, and Intersection-Over-Union).

Advanced electromagnetic sensors, nondestructive testing (NDT), and structural health monitoring (SHM) are introduced. Several electromagnetic and multi-physics sensing and imaging systems will be presented and discussed. Emerging eddy current sensors with integrated signal conditioning and digitization will be showcased. Novel resonance eddy current sensing using Inductance-to-digital converters (LDC) chips and UHF RFID sensors for corrosion monitoring under coatings, as well as EM-induced thermography, portable lamp-pulsed thermography and their applications in pipelines and railways, will be covered. Challenges and opportunities in electromagnetic NDT and SHM for condition-based operation and maintenance will be explored. Additionally, future perspectives and various applications using advanced EM sensors, NDT and SHM will be highlighted.

Recent advances in microfabrication and biotechnology have enabled the development of a wide variety of miniaturized implantable systems. Their applications range from monitoring and diagnosis to localized treatments. All these systems include, among other things, electronic components and require an energy source to power them. Although a battery can always be used, replacing it (and, more generally, the whole system) is not a routine or straightforward procedure, particularly in the case of a leadless pacemaker that is implanted directly into the patient's heart chamber. In this case, energy harvesting is a promising alternative to traditional batteries, making the systems autonomous for longer than using batteries, and enabling further miniaturization and enhanced functionality. Biomechanical energy is particularly intense in the heart region, and has the specificity of being permanently present, unlike in other parts of the human body. Thus, converting a small fraction of available mechanical energy into electricity to power a leadless pacemaker seems an ideal solution. However, the constraints in terms of dimensions, power density, reliability and durability specific to this application present unprecedented challenges. In this talk, we will present a summary of the state-of-the-art means currently being studied to meet these challenges. We'll then focus on the MEMS devices we've studied for this purpose: piezoelectric micro-cantilevers, piezoelectric microspirals, 3D electrostatic microtransducers and Silicon-on-Glass electrostatic MEMS. Finally, we will present the accelerated ageing method we have developed to assess the durability of piezoelectric energy harvesting devices, enabling to reproduce in just a few months what a piezoelectric device is mechanically subjected to an operating life of 20 years, corresponding to around 600 million heartbeats.

The climate is changing due to impact of the pollution during for several decades triggering drought conditions in many parts of the world. The agricultural sector is the most affected field of the industry and requires effective solutions to be adapted for the new conditions. Continuous periods without water supply can result the formation of cavitation bubbles inside the xylem vessels, the tension in the xylem vessels increases and result in ultrasonic emissions in the range of 100 kHz to 500 kHz. This work represents the early stage of the MEMS sensor to passively detect acoustic emissions.

Photoacoustic is an interesting medical imaging technique enabling non-invasive, high-resolution structural and functional visualization of tissues, complementary to classic ultrasonic imaging. Technological efforts are conducted to miniaturize the photoacoustic sensor systems, and make them part of the next-generation portable MEMS-based ultrasonic imaging probes. The first part of this talk will give an overview of piezoelectric micromachined ultrasonic transducers (PMUT) for receiving photoacoustically generated ultrasonic emissions, including typical materials, technology stacks, individual receiver element and array design, and typical performance. In the second part of the talk, research in progress on early designs of liquid-coupled ultrasonic piezoelectric circular membrane MEMS resonators in PiezoMUMPS SOI technology is described. Current work on assessing sensitivity of different AlN electrode geometries by FEM simulations is shown.

Chip encapsulation generates significant static mechanical stress in ICs, which affects their performance. To compensate for the offsets caused by the stress, chip designers must be able to simulate the effects of stress on their circuits and develop stress sensors and active stress-compensating circuits. To achieve this, a transistor model is presented that works from saturation to weak inversion. A method for measuring the effects of stress is presented to validate and calibrate the developed models.