INSIA LAB - News!

Radar-Based Sleep Posture Transition Sensing: An Efficient Deep Learning Hierarchical Temporal Architecture Utilizing Depthwise Convolution
(IEEE Sensors Letter, Q1, IF: 2.4)

Sleep posture transitions (SPTs) are critical indicators of sleep quality, essential for preventing posture-related complications in clinical and assistive care. Conventional vision-based, infrared, and pressure-mat monitoring systems face limitations in privacy preservation and robustness. This letter presents a sensor-oriented radar signal processing method for SPT recognition using frequency-modulated continuous-wave (FMCW) radar, processing time-range (TR) maps through a hierarchical temporal architecture. The proposed single-branch network employs depthwise convolutions, multi-head self-attention, generalized mean pooling, and bidirectional recurrent modeling to extract discriminative features from radar-signal-derived TR representations. Evaluated on a seven-class SPT dataset with 1,407 TR samples from 20 participants, the method achieves 99.29% accuracy under five-fold cross-validation. A subject-independent leave-one subject-out evaluation further confirms strong generalization, with performance of 99.22% ± 1.78% across subjects. Robustness analysis under Gaussian noise perturbation maintains 98.94% accuracy, indicating resilience to signal degradation and environmental variations. Comparative analysis with time-Doppler (TD) maps with 1,400 samples achieves 95.71% accuracy, demonstrating the superiority of range-based displacement representations over velocity based signatures. The compact 18.6M parameter architecture with 2.10 GFLOPs and 5.787 ms inference time enables real-time deployment in privacy-preserving healthcare monitoring systems while maintaining sensor-level signal fidelity. . [Details...]

Stacking Multivariate Time–Frequency Information: A Deep Learning Approach for IMU–MMG-Based Hand Gesture Recognition
(IEEE Sensors Journal, Q1, IF: 4.5)

Natural and reliable hand gesture recognition is becoming a key enabler for intuitive human-computer interaction in internet of things (IoT) environments, especially for wearable and assistive systems where camera-based sensing is unsuitable. This paper presents an inertial measurement unit (IMU) based framework that converts raw tri-axial accelerometer and gyroscope signals into structured time-frequency images and classifies them using a lightweight convolutional network. First, we propose a stacking-based multivariate time-frequency representation method termed TFR-StackGrid. In this method, the smoothed pseudo Wigner-Ville distribution (SPWVD) is computed for each of the six IMU-MMG channels. The resulting time-frequency maps are then systematically arranged into a fixed 2×3 kinematic grid. This structured configuration preserves the spatial relationship between translational acceleration and rotational motion, thereby integrating multivariate information into a single stacked time-frequency representation, referred to as S-TFR. A compact Gesture Net architecture is then designed to operate directly on these images, combining stacked 2D convolution blocks with a shallow classification head for 11 dynamic hand gesture classes. We conduct a comprehensive evaluation on the large-scale hand gesture accelerometer and gyroscope Dataset (HGAG-DATA), containing 11 gestures, 23,650 samples from 43 participants, and compare seven different time-frequency schemes, including wavelet and other quadratic distributions. The proposed SPWVD-based TFR-StackGrid with Gesture Net achieves 96.21% micro accuracy and 96.28% macro accuracy, with balanced per-class precision, recall, and F1-scores, while relying only on a single wrist-worn IMU based mechanomyography signals. These results demonstrate that structured time-frequency imaging combined with a lightweight convolutional backbone can effectively mitigate the limitations of template matching, handcrafted feature pipelines, and heavier deep sequence models for practical IMU-MMG-based hand gesture recognition. [Details...]

HatSeqNet: A Sequence-based Deep Learning Model for Cross-Subject Driver Drowsiness Detection using Single-Channel EEG Signal
(IEEE Sensors Journal, Q1, IF: 4.5)

Driver drowsiness represents a significant risk to road safety, being a major contributor to traffic accidents. This study introduces hybrid adaptive temporal sequential network HatSeqNet, an effectively designed architecture for cross-subject driver drowsiness detection using single-channel electroencephalogram (EEG) signals across different subjects. Leveraging the Mexican hat wavelet transform, EEG signals are transformed into time-scale representations (TSRs), capturing intricate neural patterns and transient dynamics. These TSRs are processed by a hybrid convolutional neural network and recurrent neural network architecture, which effectively models spatial and temporal dependencies. The model is evaluated using leave-one-subject-out cross-validation on a publicly available EEG dataset from a sustained-attention driving task. HatSeqNet achieves a state-of-the-art average accuracy of 88.26%, with precision, recall, and F1-score of 88.87%, 87.63%, and 88.10%, respectively. These metrics surpass those of prior studies on the same SAT dataset. The employment of single-channel EEG enhances the system’s portability, rendering HatSeqNet suitable for practical vehicular applications. The superior performance of the proposed architecture underscores its potential to enhance road safety through reliable and non-invasive driver drowsiness detection, even without personalized calibration. [Details...]

Improved FMCW Radar Sensing for HAR Systems: Applying Synchroextracting Transform-Enhanced Time–Frequency Representation and Custom Anisotropic CNN
(IEEE Sensors Letter, Q2, IF: 2.4)

Radar-based human activity recognition (HAR) offers non contact sensing, environmental robustness, and privacy preservation, but its performance is limited by the quality of nonstationary micro-Doppler signatures in frequency-modulated continuous-wave (FMCW) radar signals. Conventional time–frequency methods, such as short-time Fourier transform and Wigner–Ville distribution, suffer from spectral blurring and cross-term interference, reducing sensing resolution. This letter introduces an improved sensor-oriented time–frequency enhancement framework for FMCW radar HAR using the synchroextracting transform (SET), which selectively extracts energy along instantaneous fre quency ridges to suppress some artifacts and enhance micro-Doppler concentration. Applied to clutter-suppressed slow-time signals, SET produces highly concentrated and interpretable time– frequency representations (TFRs), directly improving radar sensing fidelity. The proposed approach is evaluated on the NJUST FMCW radar dataset, where SET-derived TFRs show improved concentration and activity separability over conventional methods. Using a lightweight time–frequency anisotropic radar network, the framework achieves 98.71% accuracy and a 96.14% F1-score with low computational cost (0.119 s per sample), demonstrating that sensor-level time–frequency enhancement leads to measurable gains in FMCW radar HAR systems for improved sensor-based performance. [Details...]

Two-Dimensional Successive Variational Mode Decomposition

2D extension of the successive variational mode decomposition (2D-SVMD) method is published, which decomposes 2D signals into band-limited intrinsic mode functions without requiring prior knowledge of the number of modes. The proposed 2D-SVMD addresses the limitations of existing 2D variational mode decomposition (2D-VMD) and 2D variational mode decomposition with total variation (2D-TV-VMD) methods, which need initialization of the number of modes and can produce redundant or mixed modes if the number is not accurately specified. 2D-SVMD decomposes in a sequential manner which enhances the convergence rate of the decomposition process and also mitigates the extraction of extraneous modes, collectively resulting in reduced computational overhead. The proposed 2D-SVMD is also evaluated on synthetic and real-world images and compared to 2D-VMD and 2D-TV-VMD, demonstrating its superior performance in achieving natural decomposition of 2D signals. [Details...]

EEG-based Unified Approach for Multiple Neurodevelopmental Disorders Detection in Children Using SVMD

Early age identification and prompt intervention play a crucial role in mitigating the severity of neurodevelopmental disorders in children. Traditional diagnostic approaches can be lengthy, but there is growing research potential in using electroencephalogram (EEG) signals to detect attention deficit hyperactivity disorder (ADHD) and intellectual developmental disorder (IDD). By recording the electrical activity of the brain, EEG has emerged as a promising technique for the early identification of these disorders. This research proposes a novel integrated method for identifying multiple neurodevelopmental disorders from the EEG signals of children. The approach combines successive multivariate variational mode decomposition (SMVMD) for analyzing multi-component non-stationary signals and a machine learning (ML)-based classifier, addressing the issue of inconsistent numbers of extracted features by introducing an energy-based feature integration approach. By integrating enhanced features from SMVMD with a K-nearest Neighbor (KNN) classifier, the unified approach successfully detects two separate neurodevelopmental disorders from normal subjects. [Details...]

Review of Automated Grasp Recognition using sEMG

Surface electromyography-based automated grasp recognition has emerged as a vital technology in various fields, such as automatic control, human-machine interfaces, prosthetics, and virtual reality. A review paper analyzing the available state-of-the-art methodologies for sEMG-based automated grasp recognition, covering sensing modalities, datasets, acquisition systems, pre-processing techniques, feature extraction, and identification systems has been published recently. The review also includes a detailed chronological analysis and comparison of grasp recognition techniques, focusing on the number of subjects and types of grasps. Additionally, the paper highlights open research issues and potential future prospects for these challenges, as well as several industry domains that can incorporate sEMG-based automated grasp recognition systems. [Details...]

Coronary Artery Disease Identification Using ECG

Coronary Artery Disease is a critical factor in various severe heart conditions. Early detection and treatment of CAD are crucial to prevent further disease progression. A novel approach utilizing improved energy estimation of electrocardiogram signals has been developed to automate the characterization of CAD conditions. The proposed method employs variational mode decomposition to overcome the challenges posed by the multi-component nature and nonlinear characteristics of ECG signals, enabling accurate energy estimation of ECG beats. Multiple statistical features extracted from the localized ECG beat regions were effectively classified with over 99.80% accuracy using an optimized ensemble classifier and 10-fold cross-validation. The findings demonstrate the effectiveness of the proposed methodology compared to contemporary approaches on identical datasets. [Details...]

EEG-based Brain-Computer Interface Development

Classifying Upper Limb Motor Imagery Tasks

Motor imagery research in brain-computer interfaces is critical for limb movement control. Also, cognitive robotics and engineering aim to develop brain-robot interfaces that enable effective collaboration between robots and humans. Advancements in neuroscience, robotics, and machine learning are expected to expand the scope of BRI applications, improving human-robot cooperation. Multiple approaches utilizing ratio of band power-based features, time-domain and band-power features to emphasize the inter-related information within EEG signals, combined with an optimized KNN and ensemble classifier, have been developed achieving better accuracy in categorizing MI tasks. [Details...]

Notices

We are actively looking for highly motivated BSc/MSc/PhD students (full-time/part-time/internship) to research on non-stationary signal processing, radar signal processing, biomedical signal processing, human-machine/computer interaction, brain-computer interfacing with application of artificial intelligence/machine learning for signal processing.

Page updated

Google Sites

Report abuse

Radar-Based Sleep Posture Transition Sensing: An Efficient Deep Learning Hierarchical Temporal Architecture Utilizing Depthwise Convolution (IEEE Sensors Letter, Q1, IF: 2.4)

Stacking Multivariate Time–Frequency Information: A Deep Learning Approach for IMU–MMG-Based Hand Gesture Recognition(IEEE Sensors Journal, Q1, IF: 4.5)

HatSeqNet: A Sequence-based Deep Learning Model for Cross-Subject Driver Drowsiness Detection using Single-Channel EEG Signal(IEEE Sensors Journal, Q1, IF: 4.5)

Improved FMCW Radar Sensing for HAR Systems: Applying Synchroextracting Transform-Enhanced Time–Frequency Representation and Custom Anisotropic CNN(IEEE Sensors Letter, Q2, IF: 2.4)