• MALSORI
• /
• no.51
• /
• pp.151-165
• /
• 2004

For educational and research purposes, a Korean speech recognition platform is designed. It is based on an object-oriented architecture and can be easily modified so that researchers can readily evaluate the performance of a recognition algorithm of interest. This platform will save development time for many who are interested in speech recognition. The platform includes the following modules: Noise reduction, end-point detection, met-frequency cepstral coefficient (MFCC) and perceptually linear prediction (PLP)-based feature extraction, hidden Markov model (HMM)-based acoustic modeling, n-gram language modeling, n-best search, and Korean language processing. The decoder of the platform can handle both lexical search trees for large vocabulary speech recognition and finite-state networks for small-to-medium vocabulary speech recognition. It performs word-dependent n-best search algorithm with a bigram language model in the first forward search stage and then extracts a word lattice and restores each lattice path with a trigram language model in the second stage.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.50-54
• /
• 2018

관계 추출을 위한 원격 지도 학습은 사람의 개입 없이 대규모 데이터를 생성할 수 있는 효율적인 방법이다. 그러나 원격 지도 학습은 노이즈 데이터 문제가 있으며, 노이즈 데이터는 두 가지 유형으로 나눌 수 있다. 첫 번째는 관계 표현 자체가 없는 문장이 연결된 경우이고, 두 번째는 관계 표현은 있는 문장이지만 다른 관계 표현도 함께 가지는 경우이다. 주로 문장의 길이가 길고 복잡한 문장에서 두 번째 노이즈 데이터 유형이 자주 발견된다. 본 연구는 두 번째 경우의 노이즈를 줄임으로써 관계 추출 모델의 성능을 향상시키기 위해 확장된 최단 의존 경로를 사용하는 CNN 기반 관계 추출 모델을 제안한다. 본 논문에서 제안한 방법의 우수성을 입증하기 위해, 한국어 위키피디아와 DBpedia 기반의 원격 지도 학습 데이터를 수집하여 평가한 결과, 본 논문에서 제안한 방법이 위 문제를 해결하는데 효과적이라는 것을 확인하였다.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.1
• /
• pp.101-107
• /
• 1988

This paper proposes a circuit extractor which extracts a netlist from the CIF input file cntaining the layout mask artwork informations. The circuit extractor extracts transistors and their interconnections, and calculates circuit parameter such as parasitic resistance and parasitic capacitance from the mask informations. When calculating the parasitic resistance, we consider the current flow path to reduce the errors caused by the resistance approximation. Similarly, we consider the coupling capacitance which has an effect on the circuit characteristics, when the parasitic capacitances are calculated. Therefore, using these parameter values as an input to circuit simulation, the circuit characteristics such as delay time can be estimated accurately. The presented circuit extraction algorithm uses a multiple storage quad tree as a data sturucture for storing and searching the 2-dimensional geometric data of mask artwork. Also, the proposed algorithm is technologically independent to work across a wide range of MOS technologies without any change in the algorihm.

• Proceedings of the IEEK Conference
• /
• 1999.11a
• /
• pp.629-632
• /
• 1999

In this paper, color coordinate system transformation based visual servo controller has been considered. Mobile robot always has a position error and an orientation error resulted from wheel slipping etc.. Even more, the errors have accumulative properties. So feedback from environments is important. In this paper, by using color model faster land mark extraction can be achieved. And the global position and the orientation of mobile robot can be known by only two land marks positions in image coordinate system. Finally, the adoption of visual information in path tracking problem makes visual servo control.

• Solar Energy
• /
• v.8 no.1
• /
• pp.5-12
• /
• 1988

One of the effective means to transfer the heat into and from the energy storage medium is thermosyphon. In this study, a two-phase closed thermosyphon with circular fins was used to extract the thermal energy stored in paraffin wax (Sunoco p-116). Heat transfer characteristics along the heat flow path were investigated as well as the overall performance. Some of the important results are as follows: (1) The temperature distribution of the wax in the radial direction was always maintained fairly uniformly; (2) Compared with bare thermosyphon, the heat transfer rate was vastly improved in the early stage of the experiment; and (3) Heat transfer coefficient between the wax and evaporating section of thermosyphon remained nearly constant during the experiment.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.3
• /
• pp.277-282
• /
• 2011

This paper describes a sensor fusion-based semantic map building which can improve the capabilities of a mobile robot in various domains including localization, path-planning and mapping. To build a semantic map, various environmental information, such as doors and cliff areas, should be extracted autonomously. Therefore, we propose a method to detect doors, cliff areas and robust visual features using a laser scanner and a vision sensor. The GHT (General Hough Transform) based recognition of door handles and the geometrical features of a door are used to detect doors. To detect the cliff area and robust visual features, the tilting laser scanner and SIFT features are used, respectively. The proposed method was verified by various experiments and showed that the robot could build a semantic map autonomously in various indoor environments.

• Proceedings of the IEEK Conference
• /
• 2000.07b
• /
• pp.591-595
• /
• 2000

In this paper color coordinate system transformation based visual servo controller has been considered Mobile robot always has a position error and an orientation error resulted from wheel slipping etc.. Even more, the errors have accumulative properties. So feedback from environments is important. In this paper by using color model faster land mark extraction can be achieved. And the global position and the orientation of mobile robot can be known by only two land mark positions in image coordinate system. Finally, the adoption of visual information in path tracking problem makes visual servo control.

• Journal of Korean Vacuum Science & Technology
• /
• v.7 no.1
• /
• pp.1-7
• /
• 2003

The pressure distributions in the test stand built for developing KSTAR NBI ion sources were obtained using a network system composed of conductance elements modeling the ion source, the neutralizer, and other beam line components. The allowable regime was defined on the coordinates of the gas supply rate to the ion source and the neutralizer, considering the proper conditions of the three critical parameters, the ion source pressure for good arc discharge, the pressure integral in the neutralizer for sufficient neutralization, and the chamber pressure for minimum neutral beam loss. The neutral beam evolution along the path from the ion source extraction grid to the calorimeter through the neutralizer, the bending magnet and the vacuum chamber was estimated for typical pressure distributions.

• Annual Conference of KIPS
• /
• 2012.04a
• /
• pp.321-323
• /
• 2012

최근 모바일 기기 보급의 확산과 관련 기술의 발전으로 인해 사용자의 편의를 제공하는 다양한 서비스들이 제공되고 있다. 이러한 서비스 중에서 대표적인 것으로 사용자의 이동 경로를 파악하고 예측하여 알맞은 위치기반서비스(Location-based Service; LBS)를 제공하는 것이다. 위치기반서비스를 제공하기 위한 가장 핵심 기술은 사용자의 이동 경로를 파악하는 것인데, 기존의 이동 경로 파악 기술은 이전의 이동 경로 자료를 기반으로 현재 이동 경로를 유추하였다. 그러나 이전의 이동 경로 자료가 점점 증가함에 따라 방대한 자료를 보관하고 가공하는데 많은 비용이 발생하는 문제점이 있다. 본 논문에서는 이동 경로를 예측하기 위한 방법으로 사용자가 일정 지점에 머무는 시간 정보, 이동 거리 그리고 다른 사용자와의 소통 정보를 활용한다. 이 정보들을 활용하여 사용자에게 의미 있는 장소를 추출하고 이를 기반으로 사용자의 이동 경로를 예측할 때, 기존 방법과 비교하여 적은 비용으로 효과적인 경로 예측을 할 수 있다.

• Annual Conference of KIPS
• /
• 2013.05a
• /
• pp.472-473
• /
• 2013

The growth in number and efficiency of smart devices such as GPS enabled smart phones and PDAs present an unparalleled opportunity for diverse areas of life. However extraction of GPS traces for provision of services demand a huge storage space as well as computation overhead. This is a challenging task especially for the applications which provide runtime services. In this paper we provide a simplified model to extract GPS traces of moving objects at runtime. Road segment partitioning and measure of deviation in angle of trajectory path is incorporated to identify the significant data points. The number of these data points is minimized by our proposed approach in an efficient manner to overwhelm the storage and computation overhead. Further, the competent reconstruction of complete itinerary based on gathered data, is also ensured by proposed method.