• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.10
• /
• pp.95-104
• /
• 1993

IPA(interpattern association) model is a method of feature extraction using a neural network. Even in the case that the reference patterns are simuklar to one another, this model can recover the reference patterns effectively. However, when the pattern whose feature pixels are lost is used as input, this model can not guarantee perfect recovery of the reference pattern. It is proposed a improved interpattern association(IPA) model for the feature extraction using neural network. The improved IPA model that combines the first interconnection weight matrix of the IPA model with the second additional weight matrix is proposed here to overcome the recovery problem of the original IPA model. The results of computer simulation and optical experiment are advanced.

• ETRI Journal
• /
• v.38 no.3
• /
• pp.487-493
• /
• 2016

Two new methods are proposed for an unsupervised adaptation of a language model (LM) with a single sentence for automatic transcription tasks. At the training phase, training documents are clustered by a method known as Latent Dirichlet allocation (LDA), and then a domain-specific LM is trained for each cluster. At the test phase, an adapted LM is presented as a linear mixture of the now trained domain-specific LMs. Unlike previous adaptation methods, the proposed methods fully utilize a trained LDA model for the estimation of weight values, which are then to be assigned to the now trained domain-specific LMs; therefore, the clustering and weight-estimation algorithms of the trained LDA model are reliable. For the continuous speech recognition benchmark tests, the proposed methods outperform other unsupervised LM adaptation methods based on latent semantic analysis, non-negative matrix factorization, and LDA with n-gram counting.

• Industrial Engineering and Management Systems
• /
• v.8 no.4
• /
• pp.247-256
• /
• 2009

In mass production industries such as steel making that have large equipment, sudden stops of production process due to machine failure can cause severe problems. To prevent such situations, machine diagnosis techniques play important roles. Many methods have been developed focusing on this subject. In this paper, we propose a method for the early detection of the failure on rotating machine, which is the most common theme in the machine failure detection field. A simplified method of calculating autocorrelation function is introduced and is utilized for ARMA model identification. Furthermore, an absolute deterioration factor such as Bicoherence is introduced. Machine diagnosis can be executed by this simplified calculation method of system parameter distance with weight. Proposed method proved to be a practical index for machine diagnosis by numerical examples.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.12 no.3
• /
• pp.91-100
• /
• 2004

This study presents the development of a new aluminum seat frame for the commercial bus. Back moment and seat belt anchorage analysis of the conventional steel seat frame was conducted as a base model. Effective aluminum section dimensions for aluminum pipe were calculated from equivalent stiffness and equivalent weight study. Back moment and seat belt anchorage strength with the developed aluminum seat frame were compared to those of the base model. Additionally, to pass the fatigue test, shape modification of side frame assembly was conducted. From this study we could reduce the weight of seat frame more than 5 kg. And the current analysis model and procedure can provide useful informations in designing a new commercial car seat and can reduce the overall design cost and time.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2001.04a
• /
• pp.823-826
• /
• 2001

A shape optimization of a rib of a bolster of a bogie frame is attempted and a dimension optimization on upper and lower plates is also carried out for the reduction of the weight of bogie frame. In addition, the dynamic characteristics of the weight reduced model are investigated by an analysis of a natural frequency and a transient analysis. The results show that the first natural frequency of an optimized model is larger than that of the lowest design value. And the results of transit analysis based on the experimental stress also show smaller value than the yield stress. Thus the optimized model attempted in this study is considered to be structurally stable and useful for the improvement of railway carriages.

• Journal of the Korean Society for Precision Engineering
• /
• v.23 no.7 s.184
• /
• pp.159-167
• /
• 2006

The impact time history of human body with K2 rifle on stand shooting posture is analyzed and compared with experimental results. Analysis model is Korean 50% tile of twenties and height, weight is 174 cm, 62 kgf respectively. The muscle and bone human model of Korean man's twenties 50% tile is applied. ADAMS program is used for kinematic analysis and human model is developed by Life Mod program. The effect of the ground support condition, grip position and human weight factor are studied. Maximum impact force of shoulder is 784 N. The horizontal displacement of K2 rifle muzzle is 2.9 cm.

• Journal of Ocean Engineering and Technology
• /
• v.18 no.1
• /
• pp.47-52
• /
• 2004

Greater holding force of an anchor is required for maintaining the position of a larger floating structure. According to the series of model tests of pile anchors with movable fluke, the square type pile anchor, with fluke, showed more than 6 times of the uplift pulling force, compared to the same type pile anchor, without fluke. This uplift force is 100 times its weight. When the water depth is more than 40m, It is difficult to install the pile anchor. For a convenient installation method, a type of manipulator is proposed for the separation of a weight and buoyancy controller, using TRIZ.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.1
• /
• pp.11-20
• /
• 2017

In this paper, we introduced linear spatial time series (space-time autoregressive and moving average model) and nonlinear spatial time series (space-time bilinear model). Also we estimated the parameters by Kalman Filter method and made comparative studies of power of forecast in the final model. We proposed several weight matrices such as equal proportion allocation, reciprocal proportion between distances, and proportion of population sizes. For applications, we collected Mumps data at Korea Center for Disease Control and Prevention from January 2001 until August 2008. We compared three approaches of weight matrices using the Mumps data. Finally, we also decided the most effective model based on sum of square forecast error.

• Journal of Ocean Engineering and Technology
• /
• v.27 no.4
• /
• pp.1-8
• /
• 2013

Many PE (pre-erection) blocks are supported by wooden, concrete, or steel supports when they are stocked in the outdoor areas of a shipyard. Their positions and numbers are planned on the basis of the workers' experience. Recently, many shipyards have been making PE blocks with various shapes and weight distributions because of the variety of ships and building technologies. Therefore, it is now necessary to deal with blocks that they have no experience with. We propose a method to conveniently and quickly evaluate the structural safety of PE block supports, without the need for special knowledge and technology related to structural analysis. This method can reduce the large number of man hours (MH) normally needed for the analysis. The three-dimensional grillage analysis is performed for a simplified grillage model of a PE block. For efficiency, the grillage model of the PE block is automatically built from its three-dimensional CAD model, and its weight is also automatically distributed on the grillage model. The integrated system has been comprehensively implemented to perform the grillage analysis for the reaction forces on block supports. This paper describes how to make a grillage model of a PE block and estimate the weight distribution of the block on this grillage model. These steps are verified by comparing the supports reaction forces to those of the 3D finite element analysis for the PE blocks that are provided by a shipyard.

• Journal of the Society of Naval Architects of Korea
• /
• v.56 no.6
• /
• pp.523-531
• /
• 2019

The low birthrate and the need for national defense reform in Korea drive the Navy to develop efficient human resource planning such as a manpower forecasting model. However, to our knowledge, there is no study exploring the manpower forecasting model for naval ships in Korea. The purpose of this paper is to develop a model for forecasting manpower demand in naval ships. Data for analyses were drawn from 19 ships in the Korean Navy. Results indicate that mission type is significantly related to the number of manpower. Specifically, battleships need the more manpower than the battle support ships. The results also showed that the weight of hull structure-engine and the weight of the weapons system significantly increased the number of manpower. However, the weight of the combat system was not significant. In addition, whereas the automation level of hull structure-engine and the automation level of weapon system was found to be negatively related to the number of manpower, the automation level of combat system was positively related to it. The model developed here contributes to an advanced human resource planning of the Korean Navy. Implications, limitations, and directions for future research are discussed.