Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00077
Zhiyu Jin, Zhuohe Tang, Jintao Yan
The problem of few-shot image classification has become a popular research area in the field of imaging. In order to improve the accuracy of few-shot image classification, many learning methods have been proposed, among which there are four main approaches: meta-based learning, data augmentation-based, migration-based learning and metric-based learning. In this paper, we propose a modified earth mover’s distance (MEMD) based on the metric learning approach, which has received much attention due to its simple structure and accurate classification. Similarly, MEMD constructs correlations between image regions, using such correlations to characterise the class of the image. MEMD generates a stream of best matches between image regions, and this stream of matches represents the similarity of the classified images. the MEMD algorithm requires the generation of feature weights for image regions, and EMD uses a mechanism of cross-referenced citations to generate uniform consultation weights. in contrast to EMD, MEMD generates inconsistent reference weights based on the similarity of regions. In dealing with the K-Shot problem, we used a learnable class prototype to characterise the class feature vectors. We conducted comprehensive experiments to validate our improved MEMD algorithm and tested it on four popular few-shot datasets. Namely: miniImageNet, tieredImageNet, Fewshot-CIFAR100 (FC100) and the CUB dataset.
少镜头图像分类问题已成为成像领域的研究热点。为了提高少拍图像分类的准确率，人们提出了许多学习方法，其中主要有四种方法:基于元的学习、基于数据增强的学习、基于迁移的学习和基于度量的学习。本文提出了一种基于度量学习方法的改进型推土机距离(MEMD)，该方法因其结构简单、分类准确而受到广泛关注。类似地，MEMD在图像区域之间构建相关性，使用这种相关性来表征图像的类别。MEMD生成图像区域之间的最佳匹配流，该匹配流表示分类图像的相似性。MEMD算法需要生成图像区域的特征权值，而EMD算法使用交叉引用机制生成统一的咨询权值。与EMD相比，MEMD基于区域的相似性产生不一致的参考权重。在处理K-Shot问题时，我们使用可学习的类原型来表征类特征向量。我们进行了全面的实验来验证我们改进的MEMD算法，并在四个流行的少量数据集上进行了测试。即:miniImageNet, tieredImageNet, few - shot- cifar100 (FC100)和CUB数据集。
{"title":"An Modified Earth Mover’s Distance for Few-Shot Image Classification","authors":"Zhiyu Jin, Zhuohe Tang, Jintao Yan","doi":"10.1109/PHM2022-London52454.2022.00077","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00077","url":null,"abstract":"The problem of few-shot image classification has become a popular research area in the field of imaging. In order to improve the accuracy of few-shot image classification, many learning methods have been proposed, among which there are four main approaches: meta-based learning, data augmentation-based, migration-based learning and metric-based learning. In this paper, we propose a modified earth mover’s distance (MEMD) based on the metric learning approach, which has received much attention due to its simple structure and accurate classification. Similarly, MEMD constructs correlations between image regions, using such correlations to characterise the class of the image. MEMD generates a stream of best matches between image regions, and this stream of matches represents the similarity of the classified images. the MEMD algorithm requires the generation of feature weights for image regions, and EMD uses a mechanism of cross-referenced citations to generate uniform consultation weights. in contrast to EMD, MEMD generates inconsistent reference weights based on the similarity of regions. In dealing with the K-Shot problem, we used a learnable class prototype to characterise the class feature vectors. We conducted comprehensive experiments to validate our improved MEMD algorithm and tested it on four popular few-shot datasets. Namely: miniImageNet, tieredImageNet, Fewshot-CIFAR100 (FC100) and the CUB dataset.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125975060","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00092
Julio Castro, Carolina Flores, Diego Gonzalez, Vanessa L. Quintero, Aramis Pérez
Many are the applications that use energy stored in lithium-ion (Li-ion) batteries to power themselves. Without a doubt this type of energy storage device has changed the way of living. However, a major concern regarding these Li-ion batteries is associated on the way how they should be treated or disposed when they reach their End of Life (EoL). Although, this rises a major consideration which is that perhaps the battery reached its EoL for its original application, but there are still many other applications where these batteries can be utilized. In this regard, developing a series of simple tests to understand how a degraded Li-ion battery can transfer its stored energy for less demanding applications becomes of utmost importance. This paper illustrates how the original battery pack of a drone can be reutilized for second life applications such as a small power bank. An easy procedure is proposed to assess the capability of a degraded cell to transfer energy for other applications where the Li-ion battery is working under less stress.
{"title":"From the Air to the Ground: An Experimental Approach to Assess LiPo Batteries for a Second Life","authors":"Julio Castro, Carolina Flores, Diego Gonzalez, Vanessa L. Quintero, Aramis Pérez","doi":"10.1109/PHM2022-London52454.2022.00092","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00092","url":null,"abstract":"Many are the applications that use energy stored in lithium-ion (Li-ion) batteries to power themselves. Without a doubt this type of energy storage device has changed the way of living. However, a major concern regarding these Li-ion batteries is associated on the way how they should be treated or disposed when they reach their End of Life (EoL). Although, this rises a major consideration which is that perhaps the battery reached its EoL for its original application, but there are still many other applications where these batteries can be utilized. In this regard, developing a series of simple tests to understand how a degraded Li-ion battery can transfer its stored energy for less demanding applications becomes of utmost importance. This paper illustrates how the original battery pack of a drone can be reutilized for second life applications such as a small power bank. An easy procedure is proposed to assess the capability of a degraded cell to transfer energy for other applications where the Li-ion battery is working under less stress.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"240 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121770135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00105
C. Liu
Safety is an important indicator to measure the performance of ship-borne helicopters, and the state monitoring of shipborne helicopters is the main way to ensure the safety of shipborne helicopters. As the flight time of the shipborne helicopter increases, many components will gradually degrade, and corresponding threshold need to be set to ensure the normal operation the shipborne helicopter. The commonly used fixed threshold method will cause false alarms due to different flight states, so it is necessary to dynamically construct the state thresholds under different flight states. Firstly, the empirical mode decomposition is used to denoise the signal, and then the extracted monitoring features of the shipborne helicopter are classified according to the stability in multiple flight states. A fixed threshold is setted by statistical characteristics for stable features. The dynamic feature selects flight parameters that are highly correlated with feature changes as indicators, and uses principal component analysis to fuse them to construct a dynamic threshold index.The proposed method is verified by actual flight data, and the result shows that the dynamic threshold index can effectively reduce the false alarm rate.
{"title":"Construction method of multi-stage degradation threshold for shipborne helicopter based on flight parameters","authors":"C. Liu","doi":"10.1109/PHM2022-London52454.2022.00105","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00105","url":null,"abstract":"Safety is an important indicator to measure the performance of ship-borne helicopters, and the state monitoring of shipborne helicopters is the main way to ensure the safety of shipborne helicopters. As the flight time of the shipborne helicopter increases, many components will gradually degrade, and corresponding threshold need to be set to ensure the normal operation the shipborne helicopter. The commonly used fixed threshold method will cause false alarms due to different flight states, so it is necessary to dynamically construct the state thresholds under different flight states. Firstly, the empirical mode decomposition is used to denoise the signal, and then the extracted monitoring features of the shipborne helicopter are classified according to the stability in multiple flight states. A fixed threshold is setted by statistical characteristics for stable features. The dynamic feature selects flight parameters that are highly correlated with feature changes as indicators, and uses principal component analysis to fuse them to construct a dynamic threshold index.The proposed method is verified by actual flight data, and the result shows that the dynamic threshold index can effectively reduce the false alarm rate.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"83 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126185948","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00021
Jingting You, Jun Liang, Datong Liu
Unmanned aerial vehicles play a critical role in both military and civilian applications, and their safety and reliability have also been paid more and more attention. UAV anomaly detection can detect and eliminate potential faults in a timely and effective manner, reducing the probability of accidents. Due to the influence of the complex and changeable operating environment, data shifting problems are inevitable in time series anomaly detection. Ignoring this issue may result in a significant drop in the accuracy of anomaly detection. Therefore, a UAV sensor data anomaly detection method based on Temporal Convolution Network (TCN) model transferring is proposed in this paper. First, the TCN model is pre-trained by using a large amount of data in the source domain. Then, parameters of the model are fine-tuned on the target domain. Finally, the threshold detection method is used to determine whether there is abnormality in the UAV sensor data. This work aims to address the multiple modes of UAV and improve the data-driven adaptivity for anomaly detection. In the experiments, the flight sensor data are used to verify the performance of the proposed model. The results show that the proposed method achieves high precision, high detection rate and low false detection rate in different domains.
{"title":"An Adaptable UAV Sensor Data Anomaly Detection Method Based on TCN Model Transferring","authors":"Jingting You, Jun Liang, Datong Liu","doi":"10.1109/PHM2022-London52454.2022.00021","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00021","url":null,"abstract":"Unmanned aerial vehicles play a critical role in both military and civilian applications, and their safety and reliability have also been paid more and more attention. UAV anomaly detection can detect and eliminate potential faults in a timely and effective manner, reducing the probability of accidents. Due to the influence of the complex and changeable operating environment, data shifting problems are inevitable in time series anomaly detection. Ignoring this issue may result in a significant drop in the accuracy of anomaly detection. Therefore, a UAV sensor data anomaly detection method based on Temporal Convolution Network (TCN) model transferring is proposed in this paper. First, the TCN model is pre-trained by using a large amount of data in the source domain. Then, parameters of the model are fine-tuned on the target domain. Finally, the threshold detection method is used to determine whether there is abnormality in the UAV sensor data. This work aims to address the multiple modes of UAV and improve the data-driven adaptivity for anomaly detection. In the experiments, the flight sensor data are used to verify the performance of the proposed model. The results show that the proposed method achieves high precision, high detection rate and low false detection rate in different domains.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125095817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00036
N. Sathappan, M. Tokhi, Zhanfang Zhao, Fang-wei Duan, G. Shirkoohi, Liam Penaluna
Radio-frequency Identification (RFID) technology is now widely used in many commercial and industrial sectors, from object tracking to personal identification. Few studies have investigated the possibility of using RFID systems for underwater monitoring operations in fluvial environments. While the technical limitations of these circumstances can be surmountable in certain cases, ad hoc studies have shown that RFID technology can work even under water. This paper presents the development of a device to collect and store data from a giant magnetoresistance (GMR) sensor for underwater corrosion monitoring using RFID. The findings show that RFID systems can be used to store data at near ranges of less than 5 mm regardless of frequency. This paper also provides an investigation into RFID, transponders, and reader classification, as well as existing applications underwater and their benefits.
{"title":"Underwater GMR sensor data storage using RFID tags","authors":"N. Sathappan, M. Tokhi, Zhanfang Zhao, Fang-wei Duan, G. Shirkoohi, Liam Penaluna","doi":"10.1109/PHM2022-London52454.2022.00036","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00036","url":null,"abstract":"Radio-frequency Identification (RFID) technology is now widely used in many commercial and industrial sectors, from object tracking to personal identification. Few studies have investigated the possibility of using RFID systems for underwater monitoring operations in fluvial environments. While the technical limitations of these circumstances can be surmountable in certain cases, ad hoc studies have shown that RFID technology can work even under water. This paper presents the development of a device to collect and store data from a giant magnetoresistance (GMR) sensor for underwater corrosion monitoring using RFID. The findings show that RFID systems can be used to store data at near ranges of less than 5 mm regardless of frequency. This paper also provides an investigation into RFID, transponders, and reader classification, as well as existing applications underwater and their benefits.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128389480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00066
Wei-Zhe Jiang, Xu-yun Fu, Keqiang Liu, S. Zhong, Bin Yu, Z. Bai
Aero-engine performance evaluation plays an important role in the operation and maintenance management of aviation engines, and it is the basis for formulating aero-engine operation and maintenance plans. To improve the accuracy of aero-engine performance evaluation, we propose to use the principal component analysis evaluation method to study the performance evaluation of the whole aero-engine and the module by combining the operating characteristics of the module, the original value, and the deviation value of the gas path parameters. By decoupling the original value and deviation value of high-dimensional gas path parameters, the direction with the largest performance difference if the module is selected as the principal component. Combined with the operating characteristics of modules to quantify the principal component evaluation indexes, the performance is comprehensively evaluated by using the regression model. To demonstrate the effectiveness of the method, we performed method validation using actual monitoring data from a CFM56-5B engine. The experimental results show that the method can accurately assess the whole engine and module block performance variation of the aero-engine.
{"title":"Aero-engine Performance Evaluation Based on Gas Path Parameters and Operating Characteristics","authors":"Wei-Zhe Jiang, Xu-yun Fu, Keqiang Liu, S. Zhong, Bin Yu, Z. Bai","doi":"10.1109/PHM2022-London52454.2022.00066","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00066","url":null,"abstract":"Aero-engine performance evaluation plays an important role in the operation and maintenance management of aviation engines, and it is the basis for formulating aero-engine operation and maintenance plans. To improve the accuracy of aero-engine performance evaluation, we propose to use the principal component analysis evaluation method to study the performance evaluation of the whole aero-engine and the module by combining the operating characteristics of the module, the original value, and the deviation value of the gas path parameters. By decoupling the original value and deviation value of high-dimensional gas path parameters, the direction with the largest performance difference if the module is selected as the principal component. Combined with the operating characteristics of modules to quantify the principal component evaluation indexes, the performance is comprehensively evaluated by using the regression model. To demonstrate the effectiveness of the method, we performed method validation using actual monitoring data from a CFM56-5B engine. The experimental results show that the method can accurately assess the whole engine and module block performance variation of the aero-engine.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127747563","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00057
Shanpeng Xia, Wenqiang Wang, Shengbo Zhou
This paper introduces the composition and working principle of CAN (Controller Area Network) bus communication system and describes several typical systems in detail. On this basis, the paper explains the application of CAN bus system in fault diagnosis. The conclusion shows that the proposed fault diagnosis logic CAN be used for vehicle fault diagnosis and analysis.
{"title":"Fault Diagnosis and Analysis of Automobile CAN Bus Communication","authors":"Shanpeng Xia, Wenqiang Wang, Shengbo Zhou","doi":"10.1109/PHM2022-London52454.2022.00057","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00057","url":null,"abstract":"This paper introduces the composition and working principle of CAN (Controller Area Network) bus communication system and describes several typical systems in detail. On this basis, the paper explains the application of CAN bus system in fault diagnosis. The conclusion shows that the proposed fault diagnosis logic CAN be used for vehicle fault diagnosis and analysis.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128660578","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00098
Yingshun Li, Wenbo Zhang, Huanhuan Sui, De-biao Wang
As the main component of the armored vehicle transmission system, the performance of the integrated transmission directly affects the overall performance of the vehicle. Due to the complex working conditions of the integrated transmission device, faults are prone to occur, so the fault diagnosis of the integrated transmission device is particularly important. The fault mechanism of the integrated transmission is complex, and it is difficult to directly determine the cause of the failure.Therefore,the algorithms for fault detection of the integrated transmission are constantly being introduced in recent years.This article will summarize these fault detection algorithms.
{"title":"The Survey of Algorithms for Comprehensive Transmission Fault Detection","authors":"Yingshun Li, Wenbo Zhang, Huanhuan Sui, De-biao Wang","doi":"10.1109/PHM2022-London52454.2022.00098","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00098","url":null,"abstract":"As the main component of the armored vehicle transmission system, the performance of the integrated transmission directly affects the overall performance of the vehicle. Due to the complex working conditions of the integrated transmission device, faults are prone to occur, so the fault diagnosis of the integrated transmission device is particularly important. The fault mechanism of the integrated transmission is complex, and it is difficult to directly determine the cause of the failure.Therefore,the algorithms for fault detection of the integrated transmission are constantly being introduced in recent years.This article will summarize these fault detection algorithms.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129833671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00084
Xunlei Chen, Erkang Li, Jun Yu Li, Siqi Yang, Siwen Zhang, Ziyi Wang
Security system is an important technical means of implementing security prevention and control, and its use in the field of security technology prevention is becoming more and more widespread in the current situation of expanding demand for security. The security systems used now primarily mainly rely on human visual judgment, which demonstrate the lack of intelligent analysis of video content. Static Pedestrian Intrusion Detection (SPID), which determines whether a pedestrian invades a target area in a static scene, is an important vision task in the field of intelligent video surveillance, and has a wide range of applications in scenarios such as intelligent security. To address the problem of static pedestrian intrusion detection data construction, this paper fully investigates the data set and provides sufficient data preparation for the study of this task. This paper proposes a multi-task deep network model based on target detection region segmentation and fast pedestrian detection to achieve accurate pedestrian intrusion determination in static scenes using the powerful nonlinear feature extraction capability of the network. To solve the real-time problem, the model proposes two mobile network optimization strategies, feature sharing and feature cropping, to reduce the computational complexity of the algorithm. Experimental results show that the proposed model achieves 83.1% accuracy and 20.4 FPS detection speed on the static pedestrian intrusion detection datasets, outperforming existing algorithms in terms of both accuracy and speed to achieve end-to-end real-time pedestrian intrusion detection.
{"title":"Research on Pedestrian Intrusion Detection in Static Scenes","authors":"Xunlei Chen, Erkang Li, Jun Yu Li, Siqi Yang, Siwen Zhang, Ziyi Wang","doi":"10.1109/PHM2022-London52454.2022.00084","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00084","url":null,"abstract":"Security system is an important technical means of implementing security prevention and control, and its use in the field of security technology prevention is becoming more and more widespread in the current situation of expanding demand for security. The security systems used now primarily mainly rely on human visual judgment, which demonstrate the lack of intelligent analysis of video content. Static Pedestrian Intrusion Detection (SPID), which determines whether a pedestrian invades a target area in a static scene, is an important vision task in the field of intelligent video surveillance, and has a wide range of applications in scenarios such as intelligent security. To address the problem of static pedestrian intrusion detection data construction, this paper fully investigates the data set and provides sufficient data preparation for the study of this task. This paper proposes a multi-task deep network model based on target detection region segmentation and fast pedestrian detection to achieve accurate pedestrian intrusion determination in static scenes using the powerful nonlinear feature extraction capability of the network. To solve the real-time problem, the model proposes two mobile network optimization strategies, feature sharing and feature cropping, to reduce the computational complexity of the algorithm. Experimental results show that the proposed model achieves 83.1% accuracy and 20.4 FPS detection speed on the static pedestrian intrusion detection datasets, outperforming existing algorithms in terms of both accuracy and speed to achieve end-to-end real-time pedestrian intrusion detection.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130020861","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-01DOI: 10.1109/PHM2022-London52454.2022.00020
Yihao Yu, Juanjuan Shi, Yang Luo, Weiguo Huang, Zhongkui Zhu
Gearbox plays an essential part in railway vehicles, aero engines, and other rotational machinery. Its health condition is significant for the reliable operation of such rotational machinery. To pave the way for the feature extraction of gearbox for its health condition monitoring, the vibration response of gearbox with defects is mandatory. The existing dynamic models of gearbox are mainly oriented on its individual components. However, in reality, the vibration measured from the gearbox housing usually couples the ones of the rolling bearing rotor (RBR) system, gear meshing and gearbox housing, as well as the effect of the elastohydrodynamic lubrication. Therefore, the dynamic model of individual component cannot reflect the actual vibration behaviour of the gearbox. To address this problem, a systematic model coupled the dynamic model of the gear mesh, pinion rolling bearing rotor (RBR) system, gear RBR system, the time-varying displacement excitation caused by a localized defect (LOD), and the effect of the elastohydrodynamic (EHD) lubrication has been established to predict the vibration response of gearbox in this paper. The various excited vibration of the gearbox housing can then be obtained for analyzing the vibration behaviour of the gearbox with the bearing LOD. Moreover, the influences of rotating speed and LOD sizes on the vibration characteristics of the gearbox housing are also analyzed. The established dynamic model is validated by experiments.
{"title":"The coupled vibration response analysis of gearbox with a bearing localized defect","authors":"Yihao Yu, Juanjuan Shi, Yang Luo, Weiguo Huang, Zhongkui Zhu","doi":"10.1109/PHM2022-London52454.2022.00020","DOIUrl":"https://doi.org/10.1109/PHM2022-London52454.2022.00020","url":null,"abstract":"Gearbox plays an essential part in railway vehicles, aero engines, and other rotational machinery. Its health condition is significant for the reliable operation of such rotational machinery. To pave the way for the feature extraction of gearbox for its health condition monitoring, the vibration response of gearbox with defects is mandatory. The existing dynamic models of gearbox are mainly oriented on its individual components. However, in reality, the vibration measured from the gearbox housing usually couples the ones of the rolling bearing rotor (RBR) system, gear meshing and gearbox housing, as well as the effect of the elastohydrodynamic lubrication. Therefore, the dynamic model of individual component cannot reflect the actual vibration behaviour of the gearbox. To address this problem, a systematic model coupled the dynamic model of the gear mesh, pinion rolling bearing rotor (RBR) system, gear RBR system, the time-varying displacement excitation caused by a localized defect (LOD), and the effect of the elastohydrodynamic (EHD) lubrication has been established to predict the vibration response of gearbox in this paper. The various excited vibration of the gearbox housing can then be obtained for analyzing the vibration behaviour of the gearbox with the bearing LOD. Moreover, the influences of rotating speed and LOD sizes on the vibration characteristics of the gearbox housing are also analyzed. The established dynamic model is validated by experiments.","PeriodicalId":269605,"journal":{"name":"2022 Prognostics and Health Management Conference (PHM-2022 London)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132660501","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：