检索结果-内蒙古大学图书馆

Toward Automated Air Leak Localization: A machine learning-Enhanced Ultrasonic and LiDAR-SLAM Framework for Industrial Environments

引用

ieee ACCESS 2025年 13卷 66492-66504页

作者： Schenck, Anthony Daems, Walter Steckel, Jan Univ Antwerp Fac Appl Engn Cosys Lab B-2000 Antwerp Belgium Flanders Make Strateg Res Ctr B-3920 Lommel Belgium

Pressurized air leaks are the source of a significant amount of energy loss in industrial installations. Automating the detection process of these air leaks can help prevent this loss of energy, as detecting and localizing them is the first necessary step in efficient maintenance. Typically ultrasonic microphone arrays are used to detect the leaks, since the escaping air causes turbulences around the leak opening. When the inspection is done manually, the operator of the ultrasonic sensor can evaluate whether or not an ultrasonic sound is originating from an air leak. However, during the automated process of pressurized air leak detection using ultrasonic microphone arrays, many interfering sources arise which lead to the detection of false-positive leak locations. While some of these locations can be easily filtered out using known information of the environment, such as the locations of the compressed air network, many others are more cumbersome to filter out. In order to efficiently suppress these spurious leak signals, we propose the usage of machine learning algorithms to perform classification of the signal sources. To this end, we trained a series of classifiers using real-world experiment data, and show that, by using the best performing classifiers in our leak detection algorithm, we are able to drastically reduce the amount of false-positive leaks, without any meaningful impact on the true-positive leak detection performance. We demonstrate the efficacy of the proposed approach using real-world experimental data gathered in an operating industrial setting.

关键词： Location awareness Acoustics Leak detection Microphone arrays Simultaneous localization and mapping Array signal processing Acoustic measurements Receivers Manuals Recording Air leak localization industrial acoustic sensing robot audition acoustic signal classifier

来源：评论

学校读者我要写书评

暂无评论

Noise Suppression Method With Low-Complexity Noise Estimation Model and Heuristic Noise-Masking Algorithm for Real-Time processing of Robot Vacuum Cleaners

引用

ieee ACCESS 2025年 13卷 789-801页

作者： Shin, Seunghyeon Kim, Minhan Jeon, Inkoo Song, Ju-Man Park, Yongjin Son, Jungkwan Lee, Seokjin Kyungpook Natl Univ Sch Elect & Elect Engn Daegu 41566 South Korea LG Elect Adv Robot Lab Seoul 07336 South Korea Kyungpook Natl Univ Sch Elect Engn Daegu 41566 South Korea

Noise suppression in a high-level noise environment using a low-complexity method is challenging. This study proposes a low-complexity noise suppression algorithm for robot vacuum cleaner processors. We collected working noise from a robot vacuum cleaner along with speech signals and developed a method to extract the desired speech signal while estimating the noise. Our approach estimates the noise in the existing signal and converts it into the desired signal. In addition, we designed a low-complexity neural network capable of operating on mobile processors. The evaluation results demonstrate that our method achieves a performance comparable to that of highly computational methods. Notably, our method maintains superior performance when the intensity of the desired signal is low, and its performance is less degraded than that of other methods. It exhibits less degradation than existing methods, and in contrast to other neural networks, it avoids generating incorrect signals. Furthermore, we simplified the neural network architecture reducing its size by approximately 25% with minimal performance loss.

关键词： Noise Robots Noise reduction Neural networks Complexity theory Collision avoidance Acoustics signal to noise ratio Program processors Spectrogram Source separation low-complexity low-SNR machine learning mask estimation mono channel neural network noise suppression robot vacuum cleaner

来源：评论

学校读者我要写书评

暂无评论

Resilience in Online Federated learning: Mitigating Model-Poisoning Attacks via Partial Sharing

引用

ieee Transactions on signal and Information processing over Networks 2025年 11卷 388-400页

作者： Lari, Ehsan Arablouei, Reza Gogineni, Vinay Chakravarthi Werner, Stefan Commonwealth Scientific and Industrial Research Organisation PullenvaleQLD4069 Australia University of Southern Denmark Applied AI and Data Science Unit The Maersk Mc-Kinney Moller Institute Faculty of Engineering Odense5230 Denmark Norwegian University of Science and Technology Department of Electronic Systems Trondheim7034 Norway Aalto University Department of Information and Communications Engineering Espoo02150 Finland

Federated learning (FL) allows training machine learning models on distributed data without compromising privacy. However, FL is vulnerable to model-poisoning attacks where malicious clients tamper with their local models to manipulate the global model. In this work, we investigate the resilience of the partial-sharing online FL (PSO-Fed) algorithm against such attacks. PSO-Fed reduces communication overhead by allowing clients to share only a fraction of their model updates with the server. We demonstrate that this partial sharing mechanism has the added advantage of enhancing PSO-Fed's robustness to model-poisoning attacks. Through theoretical analysis, we show that PSO-Fed maintains convergence even under Byzantine attacks, where malicious clients inject noise into their updates. Furthermore, we derive a formula for PSO-Fed's mean square error, considering factors like stepsize, attack probability, and the number of malicious clients. Interestingly, we find a non-trivial optimal stepsize that maximizes PSO-Fed's resistance to these attacks. Extensive numerical experiments confirm our theoretical findings and showcase PSO-Fed's superior performance against model-poisoning attacks compared to other leading FL algorithms. © 2015 ieee.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Near-Field Spot Beamfocusing: A Correlation-Aware Transfer learning Approach

引用

ieee TRANSACTIONS ON MOBILE COMPUTING 2025年第5期24卷 3935-3949页

作者： Fallah, Mohammad Amir Monemi, Mehdi Rasti, Mehdi Latva-aho, Matti Payame Noor Univ PNU Dept Engn Tehran 19569 Iran Univ Oulu Ctr Wireless Commun CWC Oulu 90570 Finland

Three-dimensional (3D) spot beamfocusing (SBF), in contrast to conventional angular-domain beamforming, concentrates radiating power within a very small volume in both radial and angular domains in the near-field zone. Recently the implementation of channel-state-information (CSI)-independent machine learning (ML)-based approaches have been developed for effective SBF using extremely large-scale programmable metasurface (ELPMs). These methods involve dividing the ELPMs into subarrays and independently training them with Deep Reinforcement learning to jointly focus the beam at the desired focal point (DFP). This paper explores near-field SBF using ELPMs, addressing challenges associated with lengthy training times resulting from independent training of subarrays. To achieve a faster CSI-independent solution, inspired by the correlation between the beamfocusing matrices of the subarrays, we leverage transfer learning techniques. First, we introduce a novel similarity criterion based on the phase distribution image (PDI) of subarray apertures. Then we devise a subarray policy propagation scheme that transfers the knowledge from trained to untrained subarrays. We further enhance learning by introducing quasi-liquid layers as a revised version of the adaptive policy reuse technique. We show through simulations that the proposed scheme improves the training speed about 5 times. Furthermore, for dynamic DFP management, we devised a DFP policy blending process, which augments the convergence rate up to 8-fold.

关键词： Training Array signal processing Antenna arrays Transmission line matrix methods Phased arrays Correlation Transfer learning Aperture antennas Wireless communication Transmitting antennas Spot beamfocusing near-field reinforcement learning transfer learning policy propagation policy blending quasi-liquid layer phase distribution image

来源：评论

学校读者我要写书评

暂无评论

Compressed Domain Invariant Adversarial Representation learning for Robust Audio Deepfake Detection

引用

ieee signal processing LETTERS 2025年 32卷 1111-1115页

作者： Yuan, Chengsheng Chen, Yifei Zhou, Zhili Xia, Zhihua Huang, Yongfeng Nanjing Univ Informat Sci & Technol Engn Res Ctr Digital Forens Sch Comp Sci Minist Educ Nanjing 210044 Peoples R China Guangzhou Univ Inst Artificial Intelligence & Blockchain Guangzhou 510006 Peoples R China Jinan Univ Coll Cyber Secur Guangzhou 510632 Peoples R China Tsinghua Univ Dept Elect Engn Beijing 100084 Peoples R China

The primary aim of audio deepfake detection (ADD) is to thwart deception arising from forged audio generated through text-to-speech or voice conversion technologies. However, encoding speech signals using diverse compression algorithms introduces discrepancies that significantly impair the performance of existing countermeasure systems. To tackle these challenges, this letter proposes a robust audio deepfake detection method based on Compressed Domain Invariant Adversarial Representation learning with Adaptive Token Pooling (DANet-ATP). This framework incorporates a Compression Codecs Discriminator (CCD) that, through adversarial learning in tandem with the backbone network, enhances the model's ability to extract more robust features across diverse compression codecs. Moreover, to efficiently prune redundant frame-level features while retaining vital spoofing cues, the letter designs a plug-and-play, parameter-free Adaptive Token Pooling module, significantly improving detection performance. Experimental results on the ASVspoof2021 DF dataset showcase the exceptional performance of the proposed model. Furthermore, a series of ablation experiments validate the validity and effectiveness of the proposed method.

关键词： Charge coupled devices Feature extraction Codecs Deepfakes Adaptation models Training Forgery Robustness Electronic mail Adversarial machine learning Adaptive token pooling anti-spoofing robustness audio deepfake detection domain adversarial learning

来源：评论

学校读者我要写书评

暂无评论

A Systematic Review of Surface Electromyography in Sarcopenia: Muscles Involved, signal processing Techniques, Significant Features, and Artificial Intelligence Approaches

引用

SENSORS 2025年第7期25卷 2122-2122页

作者： Leone, Alessandro Carluccio, Anna Maria Caroppo, Andrea Manni, Andrea Rescio, Gabriele Natl Res Council Italy Inst Microelect & Microsyst I-73100 Lecce Italy

Sarcopenia, affecting between 1-29% of the older population, is characterized by an age-related loss of skeletal muscle mass and function. Reduced muscle strength, either in terms of quantity or quality, and poor physical performance are among the criteria used to diagnose it. The current gold standard methods to evaluate sarcopenia are limited in terms of their cost, required expertise, and portability. A possible alternative for sarcopenia detection and monitoring is surface electromyography, which offers comprehensive information on muscle function, but a systematic synthesis of the existing literature is lacking. This systematic review aims to evaluate the application of sEMG in diagnosing and monitoring sarcopenia, focusing on the muscles involved, signal processing techniques, artificial intelligence models, and statistical analysis methods used for data interpretation. Following PRISMA guidelines, a search was performed in PubMed, Scopus, and ieee databases from 2014 up to December 2024. Original studies using sEMG for sarcopenia diagnosis or assessment in older populations were included. After removing duplicates, 145 articles were identified, of which 18 were included in the final analysis. The findings indicate a growing interest in the adoption of sEMG in sarcopenia assessment. However, methodological heterogeneity among studies limits comparability. sEMG represents a promising option for the early detection of sarcopenia, but standardized guidelines for data collection and interpretation are needed. Future studies should focus on clinical validation and results reproducibility.

关键词： surface electromyography sarcopenia wearable sensors ageing population muscles quality machine learning deep learning

来源：评论

学校读者我要写书评

暂无评论

Advances in Biomedical Missing Data Imputation: A Survey

引用

ieee ACCESS 2025年 13卷 16918-16932页

作者： Barrabes, Miriam Perera, Maria Novelle Moriano, Victor Giro-I-Nieto, Xavier Mas Montserrat, Daniel Ioannidis, Alexander G. Univ Politecn Cataluna Dept Signal Theory & Commun Barcelona 08034 Spain Stanford Univ Dept Biomed Data Sci Stanford CA 94305 USA Univ Calif Santa Cruz Dept Biomol Engn Santa Cruz CA 95060 USA

Ensuring data quality in biomedical sciences is crucial for reliable research outcomes, particularly as precision medicine continues to gain prominence. Missing values compromise data quality and can make it difficult to perform data-based studies. The origins of missing values in biomedical datasets are diverse, including experimental errors, equipment malfunctions, and variations in data collection protocols tailored to individual patient conditions. To address the complex nature of missing values and the unique characteristics of biomedical data, a diverse spectrum of computational imputation techniques has emerged. These methods range from traditional statistical analysis to more modern approaches such as discriminative machine learning models and deep generative networks. This survey paper provides a comprehensive overview of the extensive literature on missing data imputation techniques, with a specific focus on applications in genomics, single-cell RNA sequencing, health records, and medical imaging. We outline the fundamental principles underlying each imputation technique and present a detailed analysis of their advantages and disadvantages, categorized by missing data patterns. To aid practitioners in method selection, we offer practical recommendations based on critical factors such as dataset size, data type, and missingness rate. By synthesizing insights from existing literature, we provide a holistic perspective on the effectiveness of various imputation methods under different biomedical contexts, thereby facilitating informed decision-making for researchers and practitioners in applying imputation techniques to biomedical data processing.

关键词： Imputation Data models Sequential analysis Computational modeling Bioinformatics Statistics Nearest neighbor methods Multivariate regression Mathematical models Genomics Biomedical data imputation data-centric AI deep learning genomics health record imputation machine learning medical imaging missing value single-cell RNA sequence

来源：评论

学校读者我要写书评

暂无评论

A Novel Multimodal Deep learning Approach With Loss Function for Detection of Sleep Apnea Events

引用

ieee ACCESS 2025年 13卷 52085-52099页

作者： Babaei, Alireza Fakhim Tanha, Jafar Balafar, Mohammad Ali Roshan, Seyedehsan Univ Tabriz Dept Elect & Comp Engn Tabriz *** Iran

Sleep apnea, a common and potentially serious sleep disorder, is characterized by repeated pauses in breathing during sleep. These pauses, known as apneas, can last from a few seconds to minutes and may occur multiple times per hour. This disorder poses significant health risks, including increased likelihoods of developing cardiovascular complications such as hypertension, heart disease, and stroke. Detecting sleep apnea accurately and efficiently presents several challenges, including variability in physiological signals among individuals and class imbalance for apnea events. In this study, we propose a novel deep-learning approach for automatic sleep apnea detection using physiological signals. Our method utilizes a convolutional neural network (CNN) with gated recurrent units (GRU) and an attention mechanism to capture both spatial and temporal information. We introduce a multi-domain feature extractor (MDFE) block with a Squeeze and Excitation (SE) block for enhanced feature representation and channel-wise attention. To promote model diversity and robustness, we further introduce a novel loss function that encourages the network's key modules to learn disparate feature representations. This approach is trained and validated on the Sleep Heart Health Study (SHHS) dataset, achieving high accuracy and robustness (AUC: 91.94, Accuracy: 83.87). We used the K-fold cross-validation technique to show the effectiveness of our method, suggesting its potential for real-time monitoring and personalized treatment strategies in sleep disorders.

关键词： Sleep apnea Feature extraction Brain modeling Accuracy Physiology Deep learning Convolutional neural networks Training machine learning Data models attention mechanism sleep disorder sleep apnea biomedical signal processing AI for healthcare

来源：评论

学校读者我要写书评

暂无评论

A Novel CNN With Sliding Window Technique for Enhanced Classification of MI-EEG Sensor Data

引用

ieee SENSORS JOURNAL 2025年第3期25卷 4777-4786页

作者： Singh, Kamal Singha, Nitin Jaswal, Gaurav Bhalaik, Swati Natl Inst Technol NIT Delhi ECE Dept New Delhi 110036 India Indian Inst Technol Mandi Mandi 175005 Himachal Prades India Jindal Global Business Sch Sonipat 131001 Haryana India

The major challenge in fully using the motor imagery (MI)-based brain-computer interface (MI-BCI) capabilities is accurately classifying the MI electroencephalography (MI-EEG) signals. Despite numerous advancements in signal processing and deep learning (DL) techniques, there is significant scope for improvement in the accuracy currently available in the state-of-the-art. This can be achieved by leveraging spatial and temporal features of MI-EEG signal. We propose SWCNet, a convolutional neural network (CNN)-based model, and integrate it with the sliding window technique to increase the accuracy. In this work, a new CNN architecture has been proposed to extract more features from data, whereas the sliding window technique enhances temporal features by augmenting the input sensor data along the temporal dimension. We have thoroughly evaluated the performance of SWCNet using subject-dependent and subject-independent approaches for four different datasets. Our analysis includes general accuracy metrics, an ablation study, a parametric sensitivity study, and a detailed classwise performance evaluation for the tongue, foot, left-hand, and right-hand movements. The proposed model achieves accuracies of 97.42%, 94.46%, 92.27%, and 90.82% for the BCI Competition IV-2a (BCIC-IV-2a), BCI Competition IV-2b (BCIC-IV-2b), High Gamma, and OpenBMI datasets, respectively. SWCNet outperforms the state-of-the-art methods with higher accuracy for all the datasets, demonstrating its superior generalizability. SWCNet holds promise in enhancing the effectiveness of BCI applications, especially in medical rehabilitation.

关键词： Brain modeling Electroencephalography Convolutional neural networks Sensors Accuracy Feature extraction Filtering Motors Data models Electrodes Brain-computer interface (BCI) convolutional neural network (CNN) deep learning (DL) electroencephalography (EEG) machine learning (ML) motor imagery (MI) sensor

来源：评论

学校读者我要写书评

暂无评论

Effect of Artifact Removal in machine learning Based Anxiety Screening Using EEG signal 4

Effect of Artifact Removal in Machine Learning Based Anxiety...

引用

4th International Conference on Robotics, Electrical and signal processing Techniques, ICREST 2025

作者： Sakib, Nazmus Islam, Kafiul Faruk, Tasnuva Independent University Bangladesh Electrical & Electronic Engineering Dhaka Bangladesh Independent University Bangladesh Dept. of Public Health Dhaka Bangladesh

ISBN: (纸本)9798331530938

Anxiety is among the most prevalent mental health disorders. An essential component of anxiety treatment is screening. The limitations of conventional screening methods can be solved by employing machine learning to identify anxiety from EEG signals. As EEG is nonstationary, artifacts can easily corrupt the signal. The artifact in EEG data needs to be eliminated to enhance signal quality and obtain accurate analysis. The impact of artifact reduction from EEG signals for the machine learningbased anxiety screening was investigated in this study. Here, the EEG data of 40 participants (aged between 18 and 25 years) was used;of these, 24 had anxiety and 16 did not. A novel algorithm for artifact removal was used to process the data. After extracting the Hjorth parameters, the SVM classifier was used. The study revealed that implementing artifact removal improves the classification accuracy of the left frontal channels by 14.67% and the right frontal channels by 8.31% along with other performance parameters. © 2025 ieee.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：