• Journal of the Korea Society of Computer and Information
• /
• v.27 no.5
• /
• pp.55-67
• /
• 2022

Sensor data can provide fault diagnosis for equipment. However, the cause analysis for fault results of equipment is not often provided. In this study, we propose an explainable convolutional neural network framework for the sensor-based time series classification model. We used sensor-based time series dataset, acquired from vehicles equipped with sensors, and the Wafer dataset, acquired from manufacturing process. Moreover, we used Cycle Signal dataset, acquired from real world mechanical equipment, and for Data augmentation methods, scaling and jittering were used to train our deep learning models. In addition, our proposed classification models are convolutional neural network based models, FCN, 1D-CNN, and ResNet, to compare evaluations for each model. Our experimental results show that the ResNet provides promising results in the context of time series classification with accuracy and F1 Score reaching 95%, improved by 3% compared to the previous study. Furthermore, we propose XAI methods, Class Activation Map and Layer Visualization, to interpret the experiment result. XAI methods can visualize the time series interval that shows important factors for sensor data classification.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.5
• /
• pp.656-661
• /
• 2022

Password with a combination of 4-digit numbers has been widely adopted for various authentication systems (such as credit card authentication, digital door lock systems and so on). However, this system could not be safe because the 4-digit password can easily be stolen by predicting it from the fingermarks on the keypad or display screen. Furthermore, due to the prolonged COVID-19 pandemic, contactless method has been preferred over contact method in authentication. This paper suggests a new password input method based on eyeblink pattern analysis in video sequence. In the proposed method, when someone stands in front of a camera, the sequence of eyeblink motions is captured (according to counting signal from 0 to 9 or 9 to 0), analyzed and encoded, producing the desired 4-digit decimal numbers. One does not need to touch something like keypad or perform an exaggerated action, which can become a very important clue for intruders to predict the password.

• Journal of the Korean Society of Industry Convergence
• /
• v.25 no.3
• /
• pp.365-378
• /
• 2022

Complete blood cell count(CBC) is a technique that counts leukocytes for each type of blood cell being analyzed. The principle is that laser is incident to ex vivo flowing leukocytes in a microcapillary tube and scattered light occurs by laser and leukocytes. By collecting the scattered light, we can count the types of cells because different cells generate different light-scattering patterns. However, the technique has an intrinsic limitation, scattering pattern is shown in a wide range region in the resulting, which makes it difficult to accurate analyze and use fluorescent dyes. To overcome this limitation, a new design of CBC with a dual laser, which irradiates with orthogonal angles for collecting quad-scattering information was proposed. Before development, the scattering difference depending on wavelength must be investigated to only catch up to the scattered signal by angles. Some studies, which focused on simple particles, have been conducted to theoretically and experimentally investigate different scatterings by wavelength. In this study, we propose an optical system for measuring scattered light and investigate a complex particle. As a result, the green laser made strong scattering signals in both the forward and side direction: 10% and 30%, respectively.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.77-91
• /
• 2022

Various monitoring systems have been implemented in civil infrastructure to ensure structural safety and integrity. In long-term monitoring, these systems generate a large amount of data, where anomalies are not unusual and can pose unique challenges for structural health monitoring applications, such as system identification and damage detection. Therefore, developing efficient techniques is quite essential to recognize the anomalies in monitoring data. In this study, several machine learning techniques are explored and implemented to detect and classify various types of data anomalies. A field dataset, which consists of one month long acceleration data obtained from a long-span cable-stayed bridge in China, is employed to examine the machine learning techniques for automated data anomaly detection. These techniques include the statistic-based pattern recognition network, spectrogram-based convolutional neural network, image-based time history convolutional neural network, image-based time-frequency hybrid convolution neural network (GoogLeNet), and proposed ensemble neural network model. The ensemble model deliberately combines different machine learning models to enhance anomaly classification performance. The results show that all these techniques can successfully detect and classify six types of data anomalies (i.e., missing, minor, outlier, square, trend, drift). Moreover, both image-based time history convolutional neural network and GoogLeNet are further investigated for the capability of autonomous online anomaly classification and found to effectively classify anomalies with decent performance. As seen in comparison with accuracy, the proposed ensemble neural network model outperforms the other three machine learning techniques. This study also evaluates the proposed ensemble neural network model to a blind test dataset. As found in the results, this ensemble model is effective for data anomaly detection and applicable for the signal characteristics changing over time.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.5
• /
• pp.749-754
• /
• 2022

High adhesion and water resistance of the skin surface are required for wearable and skin-attachable electronic patches in various medical applications. In this study, we report a stretchable electronic patch that mimics the drainable structure pattern of the hexagonal channels of frog's pads and the sucker of an octopus based on carbon-based conductive polymer composite materials. The hexagonal channel structure that mimics the pads of frogs drains water and improves adhesion through crack arresting effect, and the suction structure that mimics an octopus sucker shows high adhesion on wet surfaces. In addition, the high-adhesive electronic patch has excellent adhesion to various surfaces such as silicone wafer (max. 4.06 N/cm²) and skin replica surface (max. 1.84 N/cm²) in dry and wet conditions. The high skin-adhesive electronic patch made of a polymer composite material based on a polymer matrix and carbon particles can reliably detect electrocardiogram (ECG) in dry and humid environments. The proposed electronic patch presents potential applications for wearable and skin-attachable electronic devices for detecting various biosignals.

• Journal of Advanced Navigation Technology
• /
• v.27 no.4
• /
• pp.457-462
• /
• 2023

We proposed a flexible link configuration in a system combining mid-span spectral inversion (MSSI) and dispersion management used for long-distance transmission of high-capacity optical signals such as wavelength division multiplexing signals, and examined specific methods to increase chromatic dispersion and nonlinear distortion compensation effects. The dispersion map proposed to increase the flexibility of dispersion-managed link configuration has a 'random-inverse' structure. That is, in the proposed dispersion map, the residual dispersion per span (RDPS) of each fiber span in the first half section up to the optical phase conjugator is randomly distributed, and the RDPS distribution in the second half section reverses the distribution pattern of the first section. Although the proposed dispersion map has a random distribution of RDPS, it was confirmed that the distortion compensation effect is improved due to the fact that the dispersion profile is symmetrical with respect to the optical phase conjugator. In the dispersion map of the 'random-inverse' configuration, it was also confirmed that the compensation effect of the distorted wavelength division multiplexing signal becomes improved when the magnitude of the RDPS allocated to each fiber span is large.

• Journal of the Korean Society for Precision Engineering
• /
• v.12 no.6
• /
• pp.13-19
• /
• 1995

A dual purpose robot automation system is developed for both arc welding and spot welding by one robot within a cell. The need for automation of both arc welding and spot welding processes is urgent while the production volume is not so big as to accommodate separate stations for the two processes. Also, space is too narrow for separate stations to be settled down in the factory. A spot welding robot is chosen and the functions for arc welding are implemented in-house at cost of advanced functions. For the spot welding, a single pole type gun is used and the robot has to push down the plate to be wolded, which causes the robot positioning error. Therefore, position error compensation algorithm is developed. The basic functions for the arc welding processes are implemented using the digital I/O board of robot controller, PLC, and A/D conversion PCB. The weaving pattern is taught in meticulously by manual teach. A fixture unit is also developed for dual purpose. The main aspects of the system is presented in this paper especially in the design and implementation procedure. The signal diagrams and sequence logic diagrams are also included. The outcome of the dual purpose welding cell is the increased productivity and good production stability which is indispensable for production volume prediction. Also, it leads to reduction of manufacturing lead time.

• Investigative Magnetic Resonance Imaging
• /
• v.5 no.1
• /
• pp.38-42
• /
• 2001

Dark band artifacts are often observed in angiograms of arteries obtained by 2D time-of-flight (TOF) angiography with saturation of veins by presaturation RF pulses. At some arteries the arterial blood velocity varies in a triphasic pattern during a cardiac cycle. The arterial blood, that is saturated by presaturation RF pulses in the saturation band, can flow back into the imaging slice during the retrograde flow phase of the triphasic variation. When such saturated retrograde flow occurs during the acquisition of the central part of the K space, a signal void can result in base images and consequently dark band artifacts can appear in angiograms. This phenomenon is experimentally demonstrated by varying the gap between the imaging slice and the saturation band. Furthermore, a new pulse sequence is proposed to eliminate the dark band artifacts by changing the profile of the saturation band front a rectangle to a ramp.

• Journal of Marine Life Science
• /
• v.3 no.2
• /
• pp.45-50
• /
• 2018

Rock bream iridovirus (RBIV) is a megalocytivirus widely infected in various fish species in Korea, causing symptoms of acute inflammation and enlargement of spleen. In our previous study, RBIV induced the initial upregulation but later down-regulation of proinflammatory cytokines and IFN1 gene expression. Signal transducers and activators of transcriptions (STAT) are transcription factors involved in the regulation of immune genes including IFNs. This study was conducted to analyse the expression of STAT2. The expressional study of STAT2 gene was performed in head kidney and spleen upon RBIV infection and immune stimulants like LPS or poly I:C in vitro. Consequently, STAT2 gene expression pattern was different in head kidney and spleen as it was significantly up-regulated by LPS from 4 h to 8 h but down-regulated at 24 h while up-regulated by poly I:C at 8 h in head kidney while, in spleen, STAT2 gene expression was down regulated by LPS but significantly up-regulated by poly I:C. Upon RBIV stimulation, STAT2 gene expression was significantly down-regulated by high dose RBIV at 4 h but up-regulated at 8 h and 24 h in head kidney. In spleen cells, it was up-regulated by medium dose RBIV at 4 h and by high dose RBIV at 4 h and 8 h but down regulated later then. In vivo, STAT2 gene expression was not significantly affected by RBIV infection while significant up-regulated by vaccination at day 7 post-vaccination, indicating STAT2 gene can be involved in adaptive immune response in rock bream.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.6
• /
• pp.305-311
• /
• 2006

In this paper the performance of a array invariant method is evaluated for source-range estimation in horizontally stratified shallow water ocean waveguide. The method has advantage of little computationally effort over existing source-localization methods. such as matched field processing or the waveguide invariant and array gain is fully exploited. And. no knowledge of the environment is required except that the received field should not be dominated by purely interference This simple and instantaneous method is applied to simulated acoustic propagation filed for testing range estimation performance. The result of range estimation according to the SNR for the underwater impulsive source with broadband spectrum is demonstrated. The spatial smoothing method is applied to suppress the effect of mutipath propagation by high frequency signal. The result of performance test for range estimation shows that the error rate is within 20% at the SNR above 10dB.