Ali Moukadem

This review provides an overview of existing accident prevention methods by monitoring the persons’ physiological state, observing movements, and physiological parameters. Firstly, different physiological parameters monitoring systems are introduced. Secondly, various systems dealing with position recognition on pressure sensing mats are presented. We conduct an in-depth literature search and quantitative analysis of papers published in this area and focus independently of the application (drivers, office and wheelchair users, etc.). Quantitative information about the number of subjects, investigated scenarios, sensor types, machine learning usage, and laboratory vs. real-world works is extracted. In posture recognition, most works recognize at least forward, backward, left and right movements on a seat. The remaining works use the pressure sensing mat for bedridden people. In physiological parameters measurement, most works detect the heart rate and often also add respiration rate recognition. Machine learning algorithms are used in most cases and are taking on an ever-greater importance for classification and regression problems. Although all solutions use different techniques, returning satisfactory results, none of them try to detect small movements, which can pose challenges in determining the optimal sensor topology and sampling frequency required to detect fine movements. For physiological measurements, there are lots of challenges to overcome in noisy environments, notably the detection of heart rate, blood pressure, and respiratory rate at very low signal-to-noise levels.

The Short Time Fourier Transform (STFT) of white noise presentwell-knownproperties regarding the distribution of its zeroes. In this letter, we propose a method to locate them beyond the usual time-frequency grid, in order to gain a better understanding on the underlying point process. To do so, we use a recent method based on off-the-grid sparse analysis, that aim at finding the best set of atomic shape forming the STFT modulus. We also improve it by ensuring the resulting combination of atoms minimizes a criterion based on the theoretical behavior of the set of zeros. Numerical experiments support our approach and show, with respect to discrete methods, an improved statistical power for the detection of signals in white noise.

A crucial step in nonintrusive load monitoring (NILM) is feature extraction, which consists of signal processing techniques to extract features from voltage and current signals. This paper presents a new time-frequency feature based on Stockwell transform. The extracted features aim to describe the shape of the current transient signal by applying an energy measure on the fundamental and the harmonic frequency voices. In order to validate the proposed methodology, classical machine learning tools are applied (k-NN and decision tree classifiers) on two existing datasets (Controlled On/Off Loads Library (COOLL) and Home Equipment Laboratory Dataset (HELD1)). The classification rates achieved are clearly higher than that for other related studies in the literature, with 99.52% and 96.92% classification rates for the COOLL and HELD1 datasets, respectively.

This paper presents the implementation of the time-frequency analysis tools, based on the S-transform and the modified S-transform, into a Raspberry Pi for a real-time computation. The aim of this approach is to provide an embedded system acquisition and measurement at low costs that would permit higher sampling resolution, recording and extraction of the data measurements in order to analyze electrical harmonics and transients. We discuss the modified S-transform proposed in literature and the different optimized Gaussian windows forms in order to obtain better energy concentration measurement. The implemented algorithm is described and applied to synthetic and real electrical signals. The modified S-transform used brings to light the enhancement of the time-frequency resolution. We conclude with future enhancements and perspectives of this study.

This paper present several signal processing tools for the analysis of heart sounds. Cardiac auscultation is noninvasive, low-cost and accurate to diagnose some heart diseases. A new module for the segmentation of heart sounds based on S-Transform is presented. The heart sound segmentation process divides the Phono Cardio Gram (PCG) signal into four parts: S1 (first heart sound), systole, S2 (second heart sound) and diastole. The segmentation can be considered one of the most important phases in the auto-analysis of PCG signals. A segmentation method based on the Shannon energy of the local spectrum calculated by the S-transform is proposed. Then, the energy concentration of the S-transform is optimized to accurately detect the boundaries of the localized sounds. New features based on the energy concentration of the S-transform are proposed to classify S1 and S2 and other features based on the complexity measure via Time-Frequency (TF) domain are proposed to detect systolic murmurs.