The regulation of tyre pressure is treated as a significant aspect of‘tyre maintenance’in the domain of autotronics.The manual supervision of a tyre pressure is typically an ignored task by most of the users.The exi...The regulation of tyre pressure is treated as a significant aspect of‘tyre maintenance’in the domain of autotronics.The manual supervision of a tyre pressure is typically an ignored task by most of the users.The existing instru-mental scheme incorporates stand-alone monitoring with pressure and/or temperature sensors and requires reg-ular manual conduct.Hence these schemes turn to be incompatible for on-board supervision and automated prediction of tyre condition.In this perspective,the Machine Learning(ML)approach acts appropriate as it exhi-bits comparison of specific performance in the past with present,intended for predicting the same in near future.The current investigation experimentally assesses the suitability of ML scheme for vibration based on-board supervision of tyre pressure of two wheeled vehicle.In order to examine the vibration response of a wheel hub,the in-house design&development of DAQ(Data Acquisition System)is described.Micro Electro-Mechanical Scheme(MEMS)built accelerometer is incorporated with open source hardware and software to collect and store the data.This framework is easy to develop,monitor and can be retrofitted in two wheeled vehicle.For various pressure conditions,the change in response of wheel hub vibration with respect to time is collected.The statistical parameters describing these vibration signals are determined and the decision tree is applied to select distinguishing parameters between extracted parameters.The classification of different conditions of tyre pressure is carried out using ML classifiers.展开更多
Domain adaptation aims to reduce the distribution gap between the training data(source domain)and the target data.This enables effective predictions even for domains not seen during training.However,most conventional ...Domain adaptation aims to reduce the distribution gap between the training data(source domain)and the target data.This enables effective predictions even for domains not seen during training.However,most conventional domain adaptation methods assume a single source domain,making them less suitable for modern deep learning settings that rely on diverse and large-scale datasets.To address this limitation,recent research has focused on Multi-Source Domain Adaptation(MSDA),which aims to learn effectively from multiple source domains.In this paper,we propose Efficient Domain Transition for Multi-source(EDTM),a novel and efficient framework designed to tackle two major challenges in existing MSDA approaches:(1)integrating knowledge across different source domains and(2)aligning label distributions between source and target domains.EDTM leverages an ensemble-based classifier expert mechanism to enhance the contribution of source domains that are more similar to the target domain.To further stabilize the learning process and improve performance,we incorporate imitation learning into the training of the target model.In addition,Maximum Classifier Discrepancy(MCD)is employed to align class-wise label distributions between the source and target domains.Experiments were conducted using Digits-Five,one of the most representative benchmark datasets for MSDA.The results show that EDTM consistently outperforms existing methods in terms of average classification accuracy.Notably,EDTM achieved significantly higher performance on target domains such as Modified National Institute of Standards and Technolog with blended background images(MNIST-M)and Street View House Numbers(SVHN)datasets,demonstrating enhanced generalization compared to baseline approaches.Furthermore,an ablation study analyzing the contribution of each loss component validated the effectiveness of the framework,highlighting the importance of each module in achieving optimal performance.展开更多
Grading of fruits based on their ripeness has been a topic of research for the last two decades.Identifying the ripened mangoes has become more of an art than science and is a challenging task.This study aims at intro...Grading of fruits based on their ripeness has been a topic of research for the last two decades.Identifying the ripened mangoes has become more of an art than science and is a challenging task.This study aims at introducing a system to grademangoes with four classes based on their ripeness.The study was demonstrated through an extensive experimentation on a newly created dataset consisting of 981 images of Alphonsomango variety belonging to four classes viz.,under-ripen,perfectly ripen,over-ripen with internal defects and over-ripen without internal defects.In this study,a hierarchical approach was adopted to classify the mangoes into the four classes.At each stage of classification,L*a*b color space features were extracted.For the purpose of classification at each stage,a number of classifiers and their possible combinationswere tried out.The study revealed that,the Support VectorMachine(SVM)classifier works better for classifyingmangoes into under-ripen,perfectly ripen and overripen while the thresholding classifier has a superior classification performance on over-ripen with internal defects and over-ripen without internal defects.Further,to bring out the superiority of the hierarchical approach,a conventional single shot multi-class classification approach with SVMwas also studied.The results of the experimentation demonstrated that the hierarchical method with an accuracy of 88%outperforms the counterpart conventional single shot multi-class classification approach in addition to several existing contemporary models.展开更多
文摘The regulation of tyre pressure is treated as a significant aspect of‘tyre maintenance’in the domain of autotronics.The manual supervision of a tyre pressure is typically an ignored task by most of the users.The existing instru-mental scheme incorporates stand-alone monitoring with pressure and/or temperature sensors and requires reg-ular manual conduct.Hence these schemes turn to be incompatible for on-board supervision and automated prediction of tyre condition.In this perspective,the Machine Learning(ML)approach acts appropriate as it exhi-bits comparison of specific performance in the past with present,intended for predicting the same in near future.The current investigation experimentally assesses the suitability of ML scheme for vibration based on-board supervision of tyre pressure of two wheeled vehicle.In order to examine the vibration response of a wheel hub,the in-house design&development of DAQ(Data Acquisition System)is described.Micro Electro-Mechanical Scheme(MEMS)built accelerometer is incorporated with open source hardware and software to collect and store the data.This framework is easy to develop,monitor and can be retrofitted in two wheeled vehicle.For various pressure conditions,the change in response of wheel hub vibration with respect to time is collected.The statistical parameters describing these vibration signals are determined and the decision tree is applied to select distinguishing parameters between extracted parameters.The classification of different conditions of tyre pressure is carried out using ML classifiers.
基金supported by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(No.RS-2024-00406320)the Institute of Information&Communica-tions Technology Planning&Evaluation(IITP)-Innovative Human Resource Development for Local Intellectualization Program Grant funded by the Korea government(MSIT)(IITP-2026-RS-2023-00259678).
文摘Domain adaptation aims to reduce the distribution gap between the training data(source domain)and the target data.This enables effective predictions even for domains not seen during training.However,most conventional domain adaptation methods assume a single source domain,making them less suitable for modern deep learning settings that rely on diverse and large-scale datasets.To address this limitation,recent research has focused on Multi-Source Domain Adaptation(MSDA),which aims to learn effectively from multiple source domains.In this paper,we propose Efficient Domain Transition for Multi-source(EDTM),a novel and efficient framework designed to tackle two major challenges in existing MSDA approaches:(1)integrating knowledge across different source domains and(2)aligning label distributions between source and target domains.EDTM leverages an ensemble-based classifier expert mechanism to enhance the contribution of source domains that are more similar to the target domain.To further stabilize the learning process and improve performance,we incorporate imitation learning into the training of the target model.In addition,Maximum Classifier Discrepancy(MCD)is employed to align class-wise label distributions between the source and target domains.Experiments were conducted using Digits-Five,one of the most representative benchmark datasets for MSDA.The results show that EDTM consistently outperforms existing methods in terms of average classification accuracy.Notably,EDTM achieved significantly higher performance on target domains such as Modified National Institute of Standards and Technolog with blended background images(MNIST-M)and Street View House Numbers(SVHN)datasets,demonstrating enhanced generalization compared to baseline approaches.Furthermore,an ablation study analyzing the contribution of each loss component validated the effectiveness of the framework,highlighting the importance of each module in achieving optimal performance.
文摘Grading of fruits based on their ripeness has been a topic of research for the last two decades.Identifying the ripened mangoes has become more of an art than science and is a challenging task.This study aims at introducing a system to grademangoes with four classes based on their ripeness.The study was demonstrated through an extensive experimentation on a newly created dataset consisting of 981 images of Alphonsomango variety belonging to four classes viz.,under-ripen,perfectly ripen,over-ripen with internal defects and over-ripen without internal defects.In this study,a hierarchical approach was adopted to classify the mangoes into the four classes.At each stage of classification,L*a*b color space features were extracted.For the purpose of classification at each stage,a number of classifiers and their possible combinationswere tried out.The study revealed that,the Support VectorMachine(SVM)classifier works better for classifyingmangoes into under-ripen,perfectly ripen and overripen while the thresholding classifier has a superior classification performance on over-ripen with internal defects and over-ripen without internal defects.Further,to bring out the superiority of the hierarchical approach,a conventional single shot multi-class classification approach with SVMwas also studied.The results of the experimentation demonstrated that the hierarchical method with an accuracy of 88%outperforms the counterpart conventional single shot multi-class classification approach in addition to several existing contemporary models.
