Measurement Engineering

Research Topic Summary.

The aim of the research is to develop a method based on energy preservation law and explore its implementation feasibility for metering instrumentation error remote detection when renewable energy sources are operating behind the sum meter and partial smart meter deployment is considered. The research group earlier investigated event-driven method for smart meter error remote estimation in grids without DERs. The hypothesis of the research is that detection of metering anomalies with the acceptable precision can be achieved by comparing measurement results over the predicted quantities according to the ML-based model built in reference conditions. The performance features of the proposed methods and the uncertainty of error estimation are evaluated using synthesized data of equivalent power networks and experimental data collected from laboratory testbeds and real power network. The expected results include techniques and models for anomaly detection in metering systems, prototypes of data collection and processing modules, scientific publications and presentations in conferences, data sets in open repositories.

Research Topic Summary.

The aim of this research is to develop an explainable artificial intelligence (XAI) framework for the automatic detection, sizing and classification of defects in aeronautical components. It should address the limitations of current AI-based non-destructive testing (NDT) systems, which often lack interpretability and struggle with complex geometries and diverse defect types. The aim is to deliver accurate, transparent, and trustworthy solutions for safety-critical aerospace applications by integrating advanced non-destructive testing with XAI techniques. The expected outcomes are an automatic defect detection, sizing and classi?cation system in aeronautic components using explainable artificial intelligence with interpretable outputs, with improved reliability in defect characterization, which should ultimately enhance inspection efficiency and ensure compliance with stringent aeronautical standards.

Research Topic Summary.

Relevance. The information technology revolution is transforming people's lives, and wearable health technologies are already widely used to monitor cardiovascular and neurological diseases. Artificial Intelligence and Machine Learning (AI/ML) techniques are being applied to the diagnosis of cardiac arrhythmias, sleep disorders, depression, epilepsy, as well as glucose and blood pressure monitoring, allowing personalised treatment. Problem. Health monitoring outside the clinic faces challenges such as uncertain recording conditions and motion artefacts that make AI/ML models unreliable. Methodologies are needed that allow not only the recording of biomarkers but also the estimation of their uncertainty, thereby increasing the diagnostic reliability of wearable technologies. Aim - to develop and investigate methods to ensure the reliability of wearable health technologies. Expected results. The developed methodologies and algorithms will enable the estimation of uncertainty in digital biomarkers, thereby increasing confidence in wearable technologies. These algorithms will be applied to the monitoring of cardiac arrhythmias, sleep disorders, and blood pressure.

Research Topic Summary.

Traumatic Brain Injury (TBI) affects 69 million people annually, with new cases ranging from 100 to 330 per 100,000 each year. TBI is a major cause of death and disability, incurring significant healthcare and societal costs in working age group. Intracranial pressure waveforms hold crucial information about brain health, with potential applications beyond TBI. This proposal explores utilizing these waveforms to enable individualized, precise medical treatments. Multimodal monitoring of intracranial waves can enhance patient outcomes, promptly identify ischemic and hyperemic events, and minimize secondary neurological damage.

Research Topic Summary.

Measurement resolution demand for reliable signals separation, but limited bandwidth is causing the signals to overlap. Research is aimed to develop the efficient signal excitation and processing techniques for ultrasonic imaging and measurements. Reliable time of flight and reflection amplitude estimation, imaging resolution and contrast are improved thanks to binary excitation spread spectrum signals application. Overlapping reflections are resolved either by spectroscopic or iterative deconvolution techniques. Correlation sidelobes, signal bandwidth can be optimized/corrected thanks to innovative excitation signals. Excitation is not limited to conventional acoustic sources, but also photoacoustics can be used. Laser ultrasound excitation can use spread spectrum signals which in turn can be adapted to spectral/correlation properties requirements. Reference signals can be made adaptable in order to increase the deconvolution efficiency. New signal quality is obtained thanks to efficient excitation and processing, which in turn enhances the possibilities for imaging and measurements.

Research Topic Summary.

The scientific problem includes the informative analysis of ultrasound signals and diagnostic images and the quantitative interpretation of the results in order to detect internal defects and pathologies. The aim is to develop and investigate the methods of analysis of informative areas in ultrasound non-destructive testing and medical diagnostic images, providing opportunities for quantitative parameter measurement and automated classification of these areas (internal defects and various pathologies), using artificial intelligence methods. The research will use advanced research infrastructure with global relevance.