• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10a
• /
• pp.1109-1114
• /
• 1992

Modern chemical processes are becoming more complicated. The sophisticated chemical processes have needed the fault diagnosis pxpert systems that can detect and diagnose the fault diagnosis expert systems that can detect and diagnose the faults of some processes and give and advice to the operator in the event of process faults. We present the Time-Delay Neural Network(TDNN) approach for on-line fautl diagnosis. The on-line fault diagnosis system finds the exact origin of the fault of which the symptom is propagated continuously with time. The proposed method has been applied to a pilot distillation column to show the merits and applicability of the TDNN.

• ETRI Journal
• /
• v.43 no.5
• /
• pp.881-890
• /
• 2021

Photoplethysmography (PPG) is a noninvasive technique that can be used to conveniently measure heart rate (HR) and thus obtain relevant health-related information. However, developing an automated PPG system is difficult, because its waveforms are susceptible to motion artifacts and between-patient variation, making its interpretation difficult. We use deep neural network (DNN) filters to mimic the cognitive ability of a human expert who can distinguish the features of PPG altered by noise from various sources. Systolic (S), onset (O), and first derivative peaks (W) are recognized by three different DNN filters. In addition, the boundaries of uninformative regions caused by artifacts are identified by two different filters. The algorithm reliably derives the HR and presents recognition scores for the S, O, and W peaks and artifacts with only a 0.7-s delay. In the evaluation using data from 11 patients obtained from PhysioNet, the algorithm yields 8643 (86.12%) reliable HR measurements from a total of 10 036 heartbeats, including some with uninformative data resulting from arrhythmias and artifacts.

• Proceedings of the KIEE Conference
• /
• 1991.07a
• /
• pp.337-340
• /
• 1991

The most difficult, time-consuming, and expensive task in building an ES (Expert System) is constructing and debugging its knowledge base. Our goals are to eliminate the knowledge-acquisition bottle-neck for ES creation in data rich situations and to make an ANN (Artificial Neural Network) model behave as much as possible like an ES. The ANN ES has many benifits: Once it has been learned, inference time is very short. It can provide a reasonable conclusion for insufficient input data. But it has also several demerits : Learning time is too long to converge. We cannot guarantee the convergence of its weights. We introduce an ANN ES model which makes most of its benefits and compensates its shortcomings.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.12a
• /
• pp.27-30
• /
• 2001

A inventory management system of the manufacturing industry has a model of different kinds according to the objective and the situation. A inventory management system needs superior system technique in demand forecast, economical efficiency, reliability and application for stable supply of the finished goods, the raw materials and the parts. This paper proposes a demand forecast method based on fuzzy structured neural network, which uses min-operation and trapezoid membership function of fuzzy rules. So we can have an intelligent inventory management system for optimized decision-making of forecasting data with expert's opinion in fuzzy environment. This inventory management system used an intelligence agent and it could be adapted to asystemenvironmentchangeinorder.

• Journal of Welding and Joining
• /
• v.25 no.6
• /
• pp.64-70
• /
• 2007

Recently, several models to control weld quality, productivity and weld properties in arc welding process have been developed and applied. Also, the applied model to make effective use of the robotic GMA(Gas Metal Arc) welding process should be given a high degree of confidence in predicting the bead dimensions to accomplish the desired mechanical properties of the weldment. In this study, a development of the on-line learning neural network models that investigate interrelationships between welding parameters and bead width as well as apply for the on-line quality control system for the robotic GMA welding process has been carried out. The developed models showed an excellent predicted results comparing with the predicted ability using off-line learning neural network. Also, the system will extend to other welding process and the rule-based expert system which can be incorporated with integration of an optimized system for the robotic welding system.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.48 no.4
• /
• pp.365-370
• /
• 1999

This paper describes the Weekly Load Forecasting Expert System(Named WLoFy) which was developed and implemented for Korea Electric Power Corporation(KEPCO). WLoFy was designed to provide user oriented features with a graphical user interface to improve the user interaction. The various forecasting models such as exponential smoothing, multiple regression, artificial nerual networks, rult-based model, and relative coefficient model also have been included in WLofy to increase the forecasting accuracy. The simulation based on historical data shows that the weekly forecasting results form WLoFy is an improvement when compared to the results from the conventional methods. Especially the forecasting accuracy on special days has been improved remakably.

• Proceedings of the KOSOMBE Conference
• /
• v.1991 no.05
• /
• pp.64-64
• /
• 1991

• Korean Journal of Cognitive Science
• /
• v.13 no.1
• /
• pp.13-21
• /
• 2002

Fuzzy logic ignores some information in the reasoning process. Neural network is powerful tools for the pattern processing, but, not appropriate for the logical reasoning. To model human knowledge, besides pattern processing capability, the logical reasoning capability is equally important. Another new neural network called neural logic network is able to do the logical reasoning. Because the fuzzy inference is a fuzzy logical reasoning, we construct fuzzy inference network based on the neural logic network, extending the existing rule-inference network. And the traditional propagation rule is modified. Experiments are performed to compare search costs by sequential searching and searching by priority. The experimental results show that the searching by priority is more efficient than the sequential searching as the size of the fuzzy inference network becomes larder and an the number of searching increases.

• Journal of Institute of Control, Robotics and Systems
• /
• v.13 no.8
• /
• pp.726-734
• /
• 2007

Generally, a mobile robot is moved by original input programs. However, it is very hard for a non-expert to change the program generating the moving path of a mobile robot, because he doesn't know almost the teaching command and operating method for driving the robot. Therefore, the teaching method with speech command for a handicapped person without hands or a non-expert without an expert knowledge to generate the path is required gradually. In this study, for easily teaching the moving path of the autonomous mobile robot, the autonomous mobile robot with the function of speech recognition is developed. The use of human voice as the teaching method provides more convenient user-interface for mobile robot. To implement the teaching function, the designed robot system is composed of three separated control modules, which are speech preprocessing module, DC servo motor control module, and main control module. In this study, we design and implement a speaker dependent isolated word recognition system for creating moving path of an autonomous mobile robot in the unknown environment. The system uses word-level Hidden Markov Models(HMM) for designated command vocabularies to control a mobile robot, and it has postprocessing by neural network according to the condition based on confidence score. As the spectral analysis method, we use a filter-bank analysis model to extract of features of the voice. The proposed word recognition system is tested using 33 Korean words for control of the mobile robot navigation, and we also evaluate the performance of navigation of a mobile robot using only voice command.

• Proceedings of the KAIS Fall Conference
• /
• 2010.11a
• /
• pp.341-343
• /
• 2010

The parameters of PID controller should be readjusted whenever system character change. In spite of a rapid development of control theory, this work needs much time and effort of expert. In this paper, to resolve this defect, after the sample of parameters in the changeable limits of system character is obtained, these parametrs are used as desired values of back propagation learning algorithm, also neural network auto tuner for PID controller is proposed by determing the optimum structure of neural network. Simulation results demonstrate that auto-tuning proper to system character can work well.