Real-time Analysis System Research Articles

Integrating intelligent algorithms in music education to analyze and improve posture and motion in instrumental training

This paper presents an innovative Artificial Intelligence (AI)—based system for real-time posture analysis and correction in instrumental music training. The system integrates OpenPose-based Convolutional Neural Networks (CNN) for skeletal tracking, Dynamic Time Warping for motion pattern analysis, and K-Nearest Neighbors (K-NN) for posture classification. Through a 16-week experimental study involving 18 music students, the system demonstrated significant improvements in learning outcomes compared to traditional methods. Key findings include (a) 33.3% faster technique acquisition in AI-assisted learning compared to traditional methods; (b) 18.6% higher posture improvement rates by week 16; (c) 40.2% better self-correction capabilities; and (d) 95.1% retention rate of correct posture after 6 months. The system processes video input at 120 fps with a total latency of 30 ms, achieving 94.3% accuracy in posture detection and 91.2% in motion pattern matching. The research establishes a comprehensive framework for integrating AI technology in music education, providing continuous, objective feedback during practice sessions. This approach addresses the critical gap between supervised instruction and individual practice, potentially reducing the risk of performance-related injuries through early detection of posture deviations.

In-vehicle edge system for real-time dashcam video analysis

In-vehicle edge system for real-time dashcam video analysis

OPTIMIZATION METHODS FOR PARAMETER IDENTIFICATION MODEL OF TEST ELECTRORETINOSIGNAL TO ASSESS NEUROTOXICITY RISKS

Introduction. The development of advanced optimization methods plays a crucial role in the enhancement of diagnostic tools in the biomedical field, particularly in the analysis of complex physiological signals. Electroretinography (ERG) is a widely used diagnostic technique that records electrical responses generated by the retina in response to light stimuli, providing valuable insights into the functional health of retinal cells. ERG is instrumental in diagnosing conditions such as retinitis pigmentosa, diabetic retinopathy, and neurotoxicity. However, the analysis of low-intensity electroretinograms (ERG) presents numerous challenges, particularly due to noise and signal distortion, which complicate accurate signal interpretation. Main purpose of this study.This paper is dedicated to developing an expert system for real-time analysis of electroretinographic signals (ERS), focusing on optimizing the parameters of a mathematical model for ERS analysis in conditions where noise and other distortions are present. The primary aim is to improve the accuracy and efficiency of ERG data processing, enabling early detection of neurotoxicity and other retinal conditions. To achieve this, we applied advanced optimization techniques, such as the Nelder-Mead method, known for its effectiveness in handling non-smooth, noisy functions. Conclusions. 1. The application of the Nelder-Mead algorithm for optimizing the complex and noisy ERS model significantly improved the performance of ERG data analysis. The algorithm's adaptability to varying optimization conditions allowed for more accurate model parameter determination, particularly in the context of real-time neurotoxicity detection. Reduction in Processing Time: The time complexity analysis revealed that the Nelder-Mead method reduced the time required to compute the model coefficients by approximately 15%. This improvement was achieved while maintaining the necessary precision for reproducing the test electroretinosignal, making it suitable for real-time applications. Computational Efficiency: One of the key findings of this study is that the use of the Nelder-Mead algorithm reduced the computational load by up to 30%. This makes the method feasible for use in expert systems designed for real-time ERS analysis, allowing for the monitoring of functional changes in the retina during the early stages of neurotoxicity detection.

Real-Time Monitoring of Metal Ions during the Molten Salt Electrolysis Process by Optical Emission Spectrometry Based on Microplasma.

Molten salt electrolysis has been widely used in the production and separation of metals, but it still lacks in situ real-time analysis methods to monitor the electrolysis process. In this work, a microplasma spectroscopic real-time analysis (MIPECA) system is developed based on noncontact direct current (DC) glow discharge. With the MIPECA system, the atomic emission spectroscopy of Li and K could be obtained in situ in LiCl-KCl molten salt, and the impact of different operating conditions on spectral signals was investigated. Then, the characteristic spectra of ten representative elements, including alkali metals (Na, Cs), alkaline earth metals (Mg, Ca, Sr, Ba), transition metals (Ag, Cu, Ni), and lanthanide metals (Ce), were all successfully excited for qualitative analysis. Under the optimal operating conditions, quantitative analysis of Ag, Cs, and Sr was achieved with high sensitivity and low limits of detection (LOD) of about 0.063, 0.021, and 0.073 wt %, respectively. Finally, the electrolysis process of Ag+ in LiCl-KCl molten salt was real-time monitored by using this MIPECA system, showing its application potential in molten salt electrolysis.

An Online Estimation Method for the Equivalent Inertia Time Constant of Power Equipment Based on Node Power Flow Equations

As renewable energy integration scales up, power systems increasingly depend on sources interfaced through power electronic converters, which lack rotating mass and substantially diminish system inertia. This reduction in inertia, coupled with the complex and diverse control strategies governing power electronics, presents significant challenges in accurately assessing the equivalent inertia levels within modern power systems. This paper introduces an online method for estimating the inertia time constant of power nodes, grounded in the node power flow equation, to address these challenges. The approach begins by deriving the rotor motion equation for synchronous generators and defining the inertia time constant of power nodes through an analysis of the power flow equations. Real-time frequency and voltage phasor data are collected from system nodes using phasor measurement units. The frequency state of the power equipment is then characterized using a divider formula, and the equivalent reactance between the power equipment and the node is further derived through the node power flow equation. This enables the real-time estimation of the equivalent inertia time constant for power nodes within the system. The effectiveness of the proposed method is demonstrated through simulations on the WSCC9 system, confirming its applicability for real-time system analysis.

Real-Time Acoustic Measurement System for Cutting-Tool Analysis During Stainless Steel Machining

This study presents a sound-based tool-wear monitoring system designed to overcome the limitations of conventional methods that focus solely on gradual and predictable wear patterns. The proposed system employs low-cost, high-frequency microphones and advanced signal processing—featuring analog/digital filtering, oversampling, signal conditioning, PLL-based synchronization, and feature extraction (ZCR, RMS)—to capture acoustic emissions during machining. Key innovations include optimized microphone placement, a custom PCB, and real-time data transfer via WiFi to MATLAB for analysis. Using the TreeBagger machine-learning algorithm, the system accurately predicts tool wear, detecting both gradual and abrupt wear patterns. Tested on EN 1.4307 (AISI/ASTM 304L) stainless steel, the system demonstrated robust performance in real-time tool-condition assessment. Its scalable and cost-effective design allows for the integration of additional sensors and features, providing a non-invasive and adaptive solution to enhance machining efficiency and reduce operational costs.

A Fast and Accurate Method for dq Impedance Modeling of Power Electronics Systems Based on Finite Differences

This paper presents a finite-difference-based method for numerically deriving the DQ impedance model of power electronics-based power systems, specifically tailored for stability analysis. The proposed method offers a computationally efficient alternative to traditional approaches by directly applying finite-difference approximations to the large-signal dynamic system, without relying on repetitive time-domain simulations or small-signal analytical models. This method eliminates the need for additional models or complex procedures to compute the steady-state solution, streamlining the impedance modeling process. The accuracy, efficiency, and precision of the proposed method are evaluated through comparative studies with analytical and time-domain perturbation methods. Results demonstrate that the proposed approach provides accuracy comparable to analytical models while significantly reducing computational effort, outperforming perturbation methods in both speed and precision. These findings highlight the practical value of the proposed method for real-time and large-scale system analysis, making it a robust tool for power systems stability assessment.

Real-Time On-Site Ground-Motion Prediction Using ResNet

Abstract On-site warnings can decrease the range of late alert zone during earthquakes. This study develops a deep learning model to predict whether the maximum peak ground acceleration at a station exceeds 25 Gal based on the waveform of the initial P-wave. A ResNet architecture model is developed to address the degradation problem in multilayered models, thereby enhancing performance. The model exhibits high performance for a test dataset, achieving F1-scores higher than 90% over various time windows, and outperforms the traditional progressive displacement value (Pdv) method and newer convolutional neural network (CNN) methods. Ground motion of two actual earthquakes are applied to test the model and validate its prediction capabilities, and the results indicate that the proposed model has higher accuracy and speed than Pdv and CNN methods. In addition, the model is integrated with the Earthworm earthquake early warning system for real-time waveform analysis and prediction, extending the model’s applicability beyond experimental stages to online systems. The proposed method is also tested using data from earthquakes that occurred in March 2023.

Optimization and Performance Analysis of Real-time Speech Translation Systems Based on Mobile Technology

As globalization deepens and mobile technology rapidly advances, the demand for crosslinguistic communication has been steadily increasing, making real-time speech translation systems a research focus. However, given the limited computational capacity and storage space of mobile devices, optimizing system performance while maintaining translation quality has become a critical challenge. Current optimization approaches for real-time speech translation systems primarily focus on improvements to model architectures and hardware acceleration, often neglecting a systematic study of model compression. This is particularly evident when handling real-time data, where achieving both high efficiency and translation accuracy remains difficult. To address these challenges, a model compression method based on the connectionist temporal classification (CTC) criterion was proposed, along with an in-depth study of parameter compression tailored for mobile applications. The research focuses on two key areas: first, model compression techniques based on the CTC criterion were explored to enhance the efficiency of real-time speech translation; second, parameter compression methods were investigated to significantly reduce resource consumption in mobile applications while preserving translation quality. The aim of this study is to improve the performance and user experience of real-time speech translation systems on mobile devices.

Structural Health Monitoring and Failure Analysis of Large-Scale Hydro-Steel Structures, Based on Multi-Sensor Information Fusion

Large-scale hydro-steel structures (LS-HSSs) are vital to hydraulic engineering, supporting critical functions such as water resource management, flood control, power generation, and navigation. However, due to prolonged exposure to severe environmental conditions and complex operational loads, these structures progressively degrade, posing increased risks over time. The absence of effective structural health monitoring (SHM) systems exacerbates these risks, as undetected damage and wear can compromise safety. This paper presents an advanced SHM framework designed to enhance the real-time monitoring and safety evaluation of LS-HSSs. The framework integrates the finite element method (FEM), multi-sensor data fusion, and Internet of Things (IoT) technologies into a closed-loop system for real-time perception, analysis, decision-making, and optimization. The system was deployed and validated at the Luhun Reservoir spillway, where it demonstrated stable and reliable performance for real-time anomaly detection and decision-making. Monitoring results over time were consistent, with stress values remaining below allowable thresholds and meeting safety standards. Specifically, stress monitoring during radial gate operations (with a current water level of 1.4 m) indicated that the dynamic stress values induced by flow vibrations at various points increased by approximately 2 MPa, with no significant impact loads. Moreover, the vibration amplitude during gate operation was below 0.03 mm, confirming the absence of critical structural damage and deformation. These results underscore the SHM system’s capacity to enhance operational safety and maintenance efficiency, highlighting its potential for broader application across water conservancy infrastructure.

Real-time modeling of transient crustal deformation through the quantification of uncertainty deduced from GNSS data

We propose a new method for real-time uncertainty monitoring of earthquake and volcano source models using data from the global navigation satellite system (GNSS) and explore its application concerning observation station placement. The Geospatial Information Authority of Japan operates two main types of GNSS earth observation network system (GEONET) coordinates for crustal deformation monitoring on different time scales: post-processing analysis values and real-time GEONET analysis system for rapid deformation monitoring (REGARD). REGARD uses the Markov Chain Monte Carlo (MCMC) method termed real-time automatic uncertainty estimation of a coseismic single rectangular model using GNSS data (RUNE) for single rectangular fault model estimation to handle uncertainty. Thus far, no GNSS monitoring system can automatically detect transient crustal deformation events, such as volcanic activity and earthquake swarms, on timescales of a day or less. We extended RUNE and developed a core program for a new monitoring system for earthquake and volcanic source models and their uncertainties. Our program achieved automatic and stable MCMC utilization for rectangular fault, dike, Mogi, and spheroid models by increasing the computational speed, improving search efficiency, and adjusting hyperparameters. The program automatically determines the standard deviation of the likelihood function assuming a normal distribution with weights for each observation station. The calculation time was within 15 s for 1 × 106 samples on a standard 1U server. We assessed the reliability of the developed method using synthetic and observed GNSS data from the 2015 Sakurajima volcanic event. The results were consistent with the assumed model and previous studies and indicated an advantage in automatically quantifying uncertainty in a short computation time. Based on MCMC samples, we developed a new visualization algorithm to indicate areas on a map in which the number of observation stations should be expanded. We assessed the reliability using data from the 2023 Noto Peninsula earthquake [Mj 6.5]. The results indicate that the algorithm is helpful in studying the placement of stations. The above model extensions and their application are essential to achieve a rapid quantitative understanding of disaster events near urban areas and for utilizing this information in emergency response activities.Graphical

DDC-Chat: Achieving accurate distracted driver classification through instruction tuning of visual language model

Driver behavior is a critical factor in road safety, highlighting the need for advanced methods in Distracted Driving Classification (DDC). In this study, we introduce DDC-Chat, a novel classification method based on a Visual large Language Model (VLM). DDC-Chat is an interactive multimodal system built upon LLAVA-Plus, fine-tuned specifically for addressing distracted driving detection. It utilizes logical reasoning chains to activate visual skills, including segmentation and pose detection, through end-to-end training. Furthermore, instruction tuning allows DDC-Chat to continuously incorporate new visual skills, enhancing its ability to classify distracted driving behavior. Our extensive experiments demonstrate that DDC-Chat achieves state-of-the-art performance on public DDC datasets, surpassing previous benchmarks. In evaluations on the 100-Driver dataset, the model exhibits superior results in both zero-shot and few-shot learning contexts, establishing it as a valuable tool for improving driving safety by accurately identifying driver distraction. Due to the computational intensity of inference, DDC-Chat is optimized for deployment on remote servers, with data streamed from in-vehicle monitoring systems for real-time analysis. More details are available on the project homepage: https://LCP-DDC-Chat.github.io/.

Integrating Diverse Data Streams for Enhanced Emotional Intelligence in Mental Health Care

This research presents a comprehensive system for real-time emotion recognition and analysis using multimodal data, including image, video, audio, and text. The system employs deep learning models to extract features and classify emotions from each modality. By integrating these predictions, we aim to provide a holistic understanding of a user's emotional state and assess potential risks. The system further analyzes the collected emotion data to identify trends, patterns, and indicators of emotional well-being and suicide risk. Visualizations such as time-series plots, distribution charts, and stacked area charts are employed to represent the emotional dynamics over time.Based on the analysis; the system generates personalized recommendations and alerts for users exhibiting signs of distress or emotional instability, including potential suicide risk factors. The goal is to promote proactive well-being management and early intervention. It is crucial to emphasize that this system is a tool for early detection and should be used in conjunction with professional mental health care.

Enhanced Diabetes Detection and Blood Glucose Prediction Using TinyML-Integrated E-Nose and Breath Analysis: A Novel Approach Combining Synthetic and Real-World Data.

Diabetes mellitus, a chronic condition affecting millions worldwide, necessitates continuous monitoring of blood glucose level (BGL). The increasing prevalence of diabetes has driven the development of non-invasive methods, such as electronic noses (e-noses), for analyzing exhaled breath and detecting biomarkers in volatile organic compounds (VOCs). Effective machine learning models require extensive patient data to ensure accurate BGL predictions, but previous studies have been limited by small sample sizes. This study addresses this limitation by employing conditional generative adversarial networks (CTGAN) to generate synthetic data from real-world tests involving 29 healthy and 29 diabetic participants, resulting in over 14,000 new synthetic samples. These data were used to validate machine learning models for diabetes detection and BGL prediction, integrated into a Tiny Machine Learning (TinyML) e-nose system for real-time analysis. The proposed models achieved an 86% accuracy in BGL identification using LightGBM (Light Gradient Boosting Machine) and a 94.14% accuracy in diabetes detection using Random Forest. These results demonstrate the efficacy of enhancing machine learning models with both real and synthetic data, particularly in non-invasive systems integrating e-noses with TinyML. This study signifies a major advancement in non-invasive diabetes monitoring, underscoring the transformative potential of TinyML-powered e-nose systems in healthcare applications.

AI-powered home cage system for real-time tracking and analysis of rodent behavior

Researchers in animal behavior and neuroscience devote considerable time to observing rodents behavior and physiological responses, with AI monitoring systems reducing personnel workload. This study presents the RodentWatch (RW) system, which leverages deep learning to automatically identify experimental animal behaviors in home cage environments. A single multifunctional camera and edge device are installed inside the animal's home cage, allowing continuous real-time monitoring of the animal's behavior, position, and body temperature for extended periods. We investigated identifying the drinking and resting behaviors of rats, with recognition accuracy enhanced through contextual object labeling and modified non-maximum suppression (NMS) schemes. Two tests-a light cycle change test and a sucrose preference test-were conducted to evaluate the usability of this system in rat behavioral experiments. This system enables notable advancements in image-based behavior recognition for living rodents.

The function of p97/valosin-containing protein (VCP) and small VCP-interacting protein (SVIP) in invasion and migration of pancreatic cancer cells.

Misfolded proteins are eliminated by a process known as endoplasmic reticulum-associated protein degradation (ERAD). ERAD has an impact on a variety of illnesses, such as diabetes, cystic fibrosis, cancer, and neurological conditions. As one of the many proteins involved in ERAD, this study is focused on p97/valosin-containing protein (VCP) and small VCP-interacting protein (SVIP). The existence and function of SVIP and p97/VCP in various types of pancreatic cancer have not yet been investigated. The study's objectives are to examine the expressions of SVIP and p97/VCP in two pancreatic cancer types and to show whether these proteins aid in the invasion and migration of pancreatic cancer cells. In this work, MIA PaCa-2 and PANC-1 human cell lines were examined. Immunocytochemistry and immunofluorescence were performed to detect the cellular localization and presence of p97/VCP and SVIP in pancreatic cancer cells. Following p97/VCP siRNA and SVIP siRNA transfection of the cells, protein expressions were assessed using Western blot analysis. The effects of this suppression on cell invasion and migration were determined using the xCELLigence real-time analysis system (RTAC). In the nucleus and cytoplasm of MIA PaCa-2 and PANC-1 cells, p97/VCP and SVIP immunoexpressions were seen. The decrease in protein expressions of p97/VCPsi and SVIPsi was significant in pancreatic cells compared to the controlsi. The p97/VCP siRNA transfection reduced the invasion and migration of MIA PaCa-2 and PANC-1 cells. In addition, the SVIP siRNA suppression resulted in increasing the invasion and migration ability of both cells. This study also demonstrated, for the first time, SVIP expression in MIA PaCa-2 and PANC-1 cells. Overall, the findings show the differential expression and function of p97/VCP and SVIP in pancreas ductal adenocarcinoma cells. The potential of the pancreatic cancer cells to migrate and invade altered when the two cell lines were transfected with p97/VCPsi and SVIPsi.

A plasmonic biosensor pre-diagnostic tool for Familial Mediterranean Fever

Familial Mediterranean Fever (FMF) is an autosomal recessive genetic disorder, primarily observed in populations around the Mediterranean Sea, linked to MEFV gene mutations. These mutations disrupt inflammatory responses, increasing pyrin-protein production. Traditional diagnosis relies on clinical symptoms, family history, acute phase reactants, and excluding similar syndromes with MEFV testing, which is expensive and often inconclusive due to heterozygous mutations. Here, we present a biosensor platform that detects differences in pyrin-protein levels between healthy and affected individuals, offering a cost-effective alternative to genetic testing. Our platform uses gold nanoparticle-based plasmonic chips enhanced with anti-pyrin antibodies, achieving a detection limit of 0.24 ng/mL with high specificity. The system integrates an optofluidic system and visible light spectroscopy for real-time analysis, with signal stability maintained for up to six months. Our technology will enhance FMF diagnosis accuracy, enabling early treatment initiation and providing a cost-effective alternative to genetic testing, thus improving patient care.

AI-RCAS: A Real-Time Artificial Intelligence Analysis System for Sustainable Fisheries Management

This study proposes an Artificial Intelligence-based Real-time Catch Analysis System (AI-RCAS) for sustainable fisheries management. The AI-RCAS, implemented on a Jetson board, consists of fish recognition using YOLOv10, tracking with a ByteTrack algorithm optimized for marine environments, and a counting module. Experiments in actual fishing environments showed significant improvements, with species recognition rates of 74–81%. The system supports the efficient operation of the total allowable catch (TAC) system through real-time analysis, addressing the limitations of the existing Electronic Monitoring (EM) systems. However, challenges remain, including object-tracking difficulties and performance issues in unstable marine environments. Future research should focus on optimizing the fishing process, improving video processing, and expanding the dataset for better generalization.

Efficiently bounding deadline miss probabilities of Markov chain real-time tasks

In real-time systems analysis, probabilistic models, particularly Markov chains, have proven effective for tasks with dependent executions. This paper improves upon an approach utilizing Gaussian emission distributions within a Markov task execution model that analyzes bounds on deadline miss probabilities for tasks in a reservation-based server. Our method distinctly addresses the issue of runtime complexity, prevalent in existing methods, by employing a state merging technique. This not only maintains computational efficiency but also retains the accuracy of the deadline-miss probability estimations to a significant degree. The efficacy of this approach is demonstrated through the timing behavior analysis of a Kalman filter controlling a Furuta pendulum, comparing the derived deadline miss probability bounds against various benchmarks, including real-time Linux server metrics. Our results confirm that the proposed method effectively upper-bounds the actual deadline miss probabilities, showcasing a significant improvement in computational efficiency without significantly sacrificing accuracy.

IoT-based real-time analysis of battery management system with long range communication and FLoRa

In the current scenario, the world is focused on renewable energy generation to achieve sustainability by 2030 regarding clean and affordable energy. Lithium-ion (Li-ion)-based Battery Energy storage (BES) is a prominent approach that is widely adopted for managing large-scale renewable energy generation. Battery Management Systems (BMS) play a critical role in optimizing battery performance of BES by monitoring parameters such as overcharging, the state of health (SoH), cell protection, real-time data, and fault detection to ensure reliability. Previous studies have concluded that the implementation of Internet of Things (IoT) with LoRa ensures effective real-time monitoring of the BMS of Li-ion batteries. This study proposed and implemented a customized LoRa and IoT-based hardware system with a gateway to acquire parameters such as terminal voltage, current, charge voltage, charge current, cycle, temperature, state of charge (SoC), and SoH, and log them into the cloud server. An OMnet++-based Framework for LoRa (FLoRa) simulation was implemented to analyze the power consumption and residual energy of the customized LoRa nodes. The simulation was configured with a spread factor of 7, a carrier frequency of 433 MHz, a bandwidth of 125 KHz, and a transmission power of 2 dBm. The simulation results indicated that Node 3 had the highest mean power consumption (0.028233) and total energy consumption (0.146592), whereas Node 0 exhibited the lowest mean power consumption (0.023413) and total energy consumption (0.070204). Additionally, a comprehensive dataset encompassing voltage, current, and time was created and utilized for precise calculations of the battery's capacity and state of health, with potential use in future predictions.