• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.127.1-127
• /
• 2001

In this paper we propose an input/output model based fault diagnosis method to detect and isolate single faults in the robot arm control system. The proposed algorithm is functionally composed of three main parts-parameter estimation, fault detection, and isolation, When a change in the system occurs, the errors between the system output and the estimated output cross a predetermined threshold, and once a fault in the system is detected, and in this zone the estimated parameters are transferred to the fault classifier by ART2(adaptive resonance theory 2) neural network for fault isolation. Since ART2 neural network is an unsupervised neural network fault classifier does not require the knowledge of all possible faults to isolate the faults occurred in the system. Simulations are carried out to evaluate the performance of the proposed ...

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.11
• /
• pp.319-322
• /
• 1997

This paper presents a CAD (Computer-Aided Diagnosis) system or detection of breast cancer, which is composed of personal computer, X-ray film scanner, high resolution display and application softwares. There are three major algorithms implemented in the application software. The irst algorithm is the adaptive enhancement of the digitized X-ray mammograms based on the first derivative and the local statistics. The second one is to detect the clustered microcalcifications by using the statistical texture analysis, and the third one is the classification of the clustered microcalcifications as malignant or benign by using the shape analysis. These algorithms were verified by real experiments.

• Journal of Intelligence and Information Systems
• /
• v.14 no.2
• /
• pp.193-206
• /
• 2008

Medical diagnosis can be considered a classification task which classifies disease types from patient's condition data represented by a set of pre-defined attributes. This study proposes a hybrid genetic algorithm based classification method to develop classifiers for multidimensional pattern classification problems related with medical decision making. The classification problem can be solved by identifying separation boundaries which distinguish the various classes in the data pattern. The proposed method fits a finite number of regional agents to the data pattern by combining genetic algorithms and local adaptive operations. The local adaptive operations of an agent include expansion, avoidance and relocation, one of which is performed according to the agent's fitness value. The classifier system has been tested with well-known medical data sets from the UCI machine learning database, showing superior performance to other methods such as the nearest neighbor, decision tree, and neural networks.

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.05
• /
• pp.414-417
• /
• 1997

This paper presents vessel contour for extracting features and segmentating narrow blood vessels down to a diameter of two pixels in digital subtraction angiographic image. We present a new tracking algorithm for contour, mainly blood vessels in DSA image, and extracting properties such as their intensities, diameters, and center lines by exploiting spatial continuity. The proposed algorithm comes to detect blood vessel's boundary using difference edge detector one of homogeneity operator and find a next centerline position by direction vector of edge information. This algorithm enhanced variation of vessel's diameter compared to Sun's tracking algorithm and lessoned to compute as direction vector decide adaptively entire vessel's direction field. The processed images are intended to support radiologists in diagnosis, radiation therapy planning, and surgical planning. The algorithm should be useful for automating angiographic analyses of blood vessels.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.10 no.4
• /
• pp.160-167
• /
• 2007

In this paper, we presented the development results of high speed wavefront sensor which is used in diagnosing the beam quality of He-Ne laser for adaptive optics system. The beam quality information of laser in AO system is necessarily required for diagnosing the optical components or correcting the distorted wavefront afterward. According to system requirements, normally, it is requested that there are high precision of measurement and real time processing speed. The developed wavefront sensor in this paper achieved maximum 30Hz of measurement rate and ${\lambda}/20(\;{@}\;{\lambda}=0.6328{\mu}m)$ of measurement precision in RMS. We also applied the developed into an experimental adaptive system and verified the performance of it by correcting the aberrated wavefront with a rate of 30Hz and $\lambda$/20 precision using the combination of the developed and PID control algorithm.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.10
• /
• pp.2223-2230
• /
• 2011

ECG is used to diagnose heart diseases such as myocardial ischemia, arrhythmia and myocardial infarction. Particularly, myocardial ischemia causes the shape change of the ST segment, this change is transient and may occur without symptoms. So it is important to detect the transient change of ST segment through long term monitoring. ST segment classification algorithm for making diagnosis myocardial ischemia is presented in this paper. The first step in the ST segment shape classification process is to detect R wave point and feature points based adaptive threshold and window. And then, the suggested algorithm detects the ST level change, To classify the ST segment shape, the suggested algorithm uses the slope values of the four points between the S and T wave. The ECG data in the European ST-T database were used to verify the performance of the developed algorithm. The best correct rate was 99.40% and the worst correct rate was 68.48%.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.40 no.6
• /
• pp.1120-1129
• /
• 2015

In this paper, we propose an algorithm to detect movement noise in PPG(Photoplethysmography) measurements. Movement noise significantly deteriorate PPG signals in measurement, so that a movement noise detection algorithm is critical before using measured PPG signals for applications such as diagnosis. To detect movement noise, we apply wavelet transform to PPG signals instead of short-time Fourier transform and decide if the measured signlas include movement noise. To that end, we adaptively choose a wavelet, which is the most similar to the subject's PPG pattern. In the case when movement noise is intentionally added in the 20% and 30% of the total experiment time, our algorithm detects time-slots including movement and outperforms previous works.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.10 no.3
• /
• pp.211-217
• /
• 2000

A main contribution of this paper is the development of a Hopfield network-based algorithm for the fault diagnosis of the actuators in linear system with uncertainties. An unknown input decoupling approach is introduced to the design of an adaptive observer so that the observer is insensitive to uncertainties. As a result, the output observation error equation does not depend on the effect of uncertainties. Simultaneous energy minimization by the Hopfield network is used to minimize the least mean square of errors of errors of estimates of output variables. The Hopfield network provides an estimate of the gains of the actuators. When the system dynamics changes, identified gains go through a transient period and this period is used to detect faults. The proposed scheme is demonstrated through its application to a simulated second-order system.

• KIEE International Transactions on Electrophysics and Applications
• /
• v.5C no.3
• /
• pp.138-142
• /
• 2005

In this paper, we compared recognition rates between NN(neural networks) and clustering method as a scheme of off-line PD(partial discharge) diagnosis which occurs at the stator coil of traction motor. To acquire PD data, three defective models are made. PD data for classification were acquired from PD detector. And then statistical distributions are calculated to classify model discharge sources. These statistical distributions were applied as input data of two classification tools, BP(Back propagation algorithm) and ANFIS(adaptive network based fuzzy inference system) pre-processed FCM(fuzzy c-means) clustering method. So, classification rate of BP were somewhat higher than ANFIS. But other items of ANFIS were better than BP; learning time, parameter number, simplicity of algorithm.

• Industrial Engineering and Management Systems
• /
• v.10 no.4
• /
• pp.272-278
• /
• 2011

Decision tree as a classification tool is being used successfully in many areas such as medical diagnosis, customer churn prediction, signal detection and so on. The main advantage of decision tree classifiers is their capability to break down a complex structure into a collection of simpler structures, thus providing a solution that is easy to interpret. Since decision tree is a top-down algorithm using a divide and conquer induction process, there is a risk of reaching a local optimal solution. This paper proposes a procedure of optimally determining thresholds of the chosen variables for a decision tree using an adaptive particle swarm optimization (APSO). The proposed algorithm consists of two phases. First, we construct a decision tree and choose the relevant variables. Second, we find the optimum thresholds simultaneously using an APSO for those selected variables. To validate the proposed algorithm, several artificial and real datasets are used. We compare our results with the original CART results and show that the proposed algorithm is promising for improving prediction accuracy.