• 대한인간공학회지
• /
• 제28권1호
• /
• pp.37-51
• /
• 2009

The nuclear power industry in the world has recognized the importance of integrating non-technical and team skills training with the technical training given to its control room operators to reduce human errors since the Three Mile Island and Chernobyl accidents. The Nuclear power plant (NPP) industry in Korea has been also making efforts to reduce the human errors which largely have contributed to 120 nuclear reactor trips from the year 2001 to 2006. The Crew Resource Management (CRM) training was one of the efforts to reduce the human errors in the nuclear power industry. The CRM was developed as a response to new insights into the causes of aircraft accidents which followed from the introduction of flight recorders and cockpit voice recorders into modern jet aircraft. The CRM first became widely used in the commercial airline industry, but military aviation, shipboard crews, medical and surgical teams, offshore oil crews, and other high-consequence, high-risk, time-critical industry teams soon followed. This study aims to develop a CRM training program that helps to improve plant performance by reducing the number of reactor trips caused by the operators' errors in Korean NPP. The program is; firstly, based on the work we conducted to develop a human factors training from the applications to the Nuclear Power Plant; secondly, based on a number of guidelines from the current practicable literature; thirdly, focused on team skills, such as leadership, situational awareness, teamwork, and communication, which have been widely known to be critical for improving the operational performance and reducing human errors in Korean NPPs; lastly, similar to the event-based training approach that many researchers have applied in other domains: aircraft, medical operations, railroads, and offshore oilrigs. We conducted an experiment to test effectiveness of the CRM training program in a condition of simulated control room also. We found that the program made the operators' attitudes and behaviors be improved positively from the experimental results. The more implications of the finding were discussed further in detail.

• 음성과학
• /
• 제15권2호
• /
• pp.55-65
• /
• 2008

Recently, the multi-model based speech recognizer has been used quite successfully for noisy speech recognition. For the selection of the reference HMM (hidden Markov model) which best matches the noise type and SNR (signal to noise ratio) of the input testing speech, the estimation of the SNR value using the VAD (voice activity detection) algorithm and the classification of the noise type based on the GMM (Gaussian mixture model) have been done separately in the multi-model framework. As the SNR estimation process is vulnerable to errors, we propose an efficient method which can classify simultaneously the SNR values and noise types. The KL (Kullback-Leibler) distance between the single Gaussian distributions for the noise signal during the training and testing is utilized for the classification. The recognition experiments have been done on the Aurora 2 database showing the usefulness of the model compensation method in the multi-model based speech recognizer. We could also see that further performance improvement was achievable by combining the probability density function of the MCT (multi-condition training) with that of the reference HMM compensated by the D-JA (data-driven Jacobian adaptation) in the multi-model based speech recognizer.

• 음성과학
• /
• 제9권4호
• /
• pp.237-252
• /
• 2002

MIT has been applied for nonfluent aphasic patients on the basis of lateralization of brain hemisphere. However, its applications for different languages have some inquiry for aphasic patients because of prosodic and rhythmic differences. The purpose of this study is to develop the Korean Melodic Intonation Therapy program using personal computer and its clinical effects for nonfluent aphasic patients. The algorithm was composed to voice analog signal, PCM, AMDF, Short-time autocorrelation function and center clipping. The main menu contains pitch, waveform, sound intensity and speech files on window. Aphasic patients' intonation patterns overlay on selected kMIT patterns. Three aphasic patients with or without kMIT training participated in this study. Four affirmative sentences and two interrogative sentences were uttered on CSL by stimulus of ST. VOT, VD, Hold and TD were measured on Spectrogram. In addition, articulation disorders and intonation patterns were evaluated objectively on spectrogram. The results indicated that nonfluent aphasic patients with kMIT training group showed some clinical effects of speech intelligibility based on VOT, TD values, articulation evaluation and prosodic pattern changes.

• 영어영문학
• /
• 제56권5호
• /
• pp.871-905
• /
• 2010

This paper is motivated by Watanabe's (2001) observation that English byphrase passives are sometimes translated into Japanese object topicalization constructions. That is, the original English sentence in the passive may be translated into the active voice with the logical object topicalized. A number of scholars, including Chomsky (1981) and Baker (1992), have remarked that languages have various ways to avoid focusing on the logical subject. The aim of the present study is to examine the translation equivalents of the English by-phrase passives in an English-Korean parallel corpus compiled by the author. A small sample of articles from Newsweek magazine and its published Korean translation reveals that there are indeed many ways to translate English by-phrase passives, including object topicalization (12.5%). Among the 64 translated sentences analyzed and classified, 12 (18.8%) examples were problematic in terms of agent defocusing, which is the primary function of passives. Of these 12 instances, five cases were identified where an alternative translation would be more suitable. The results suggest that the functional characteristics of English by-phrase passives should be highlighted in translator training as well as language teaching.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2015년도 추계학술대회
• /
• pp.167-170
• /
• 2015

본 논문에서는 바쁜 현대인들뿐만 아니라 노인 및 어린이들이 보다 편안하고 효율적으로 애완동물을 기를 수 있도록 하기 위한 애완동물 배뇨 훈련 및 먹이 자동 공급을 위한 시스템을 제안한다. 이 시스템은 블루투스를 이용해 스마트폰 어플리케이션으로 조작할 수 있도록 하며, 나아가 Wifi를 통해 인터넷에 접속하여 어디서든 조작할 수 있도록 한다. 본 논문의 시스템은 배뇨판과 먹이 공급기로 나뉘어져있다. 배뇨판은 배뇨 인식을 위한 수압 센서와 음성출력을 위한 녹음기 모듈, 스피커로 구성되고 후면부는 먹이 자동 공급을 위한 두 개의 서보 모터와 원격 통신을 위한 블루투스 센서로 구성된다. 배뇨판과 먹이 공급기 모두 아두이노 보드와 C언어 기반의 아두이노 스케치 프로그램으로 제작하였으며 먹이 공급기는 블루투스 통신을 지원하는 라이브러리로 통신할 수 있게 하였다. 구현된 시스템은 애완동물의 종류 및 크기에 관계없이 자동 수위조절과 먹이양이 조절되는 특징을 가진다.

• 한국컴퓨터정보학회논문지
• /
• 제26권9호
• /
• pp.57-64
• /
• 2021

본 논문에서는 음성 명령을 인식하여 비행기의 1차 조종면을 제어할 수 있는 장치를 제안한다. 음성 명령어는 19개의 명령어로 구성되며 총 2,500개의 데이터셋을 근간으로 학습 모델을 구성한다. 학습 모델은 TensorFlow 기반의 Keras 모델의 Sequential 라이브러리를 이용하여 CNN 모델로 구성되며, 학습에 사용되는 음성 파일은 MFCC 알고리즘을 이용하여 특징을 추출한다. 특징을 인식하기 위한 2단계의 Convolution layer 와 분류를 위한 Fully Connected layer는 2개의 dense 층으로 구성하였다. 검증 데이터셋의 정확도는 98.4%이며 테스트 데이터셋의 성능평가에서는 97.6%의 정확도를 보였다. 또한, 라즈베리 파이 기반의 제어장치를 설계 및 구현하여 동작이 정상적으로 이루어짐을 확인하였다. 향후, 음성인식 자동 비행 및 항공정비 분야의 가상 훈련환경으로 활용될 수 있을 것이다.

• 지적과 국토정보
• /
• 제52권1호
• /
• pp.81-93
• /
• 2022

최근 자율주행 기술에서 차량 주변 객체 인식과 교통표지판 및 차량 신호 인식을 위한 연구가 활발히 수행되고 있으며, 특히 차량 신호 인식은 자율주행 기술에 있어서 핵심 요소로 평가되고 있다. 이에 차량 신호 인식을 위한 다양한 연구가 진행되어 왔으며, 최근에는 딥러닝 기반 객체 인식 모델을 활용한 차량 신호 인식 연구가 크게 증가하고 있다. 또한 AIHub에서 음성, 비전, 자율주행 등을 위한 양질의 국내 인공지능 학습데이터 셋이 공개됨에 따라 이들 데이터를 활용한 국내 환경에 적합한 차량 신호 인식 모델의 개발도 가능하게 되었다. 이에 본 연구에서는 AIHub의 학습데이터와 객체 인식모델 YOLO를 적용한 국내 차량 신호 인식 모델을 개발하였다. 특히 차량 신호의 인식 성능을 개선하기 위하여 YOLOv4와 YOLOv5의 다양한 모델을 적용하였으며 학습데이터의 클래스도 다양하게 분류하여 실험을 수행하였다. 결론적으로 YOLOv5가 YOLOv4보다 차량 신호 인식에 조금 더 적합함을 확인할 수 있었으며, 두 모델의 아키텍처 비교를 통하여 YOLOv5 성능이 우수한 이유를 확인할 수 있었다.

• 대한음성학회지:말소리
• /
• 제52호
• /
• pp.161-169
• /
• 2004

This paper describes the automated call routing for train ticket reservation aid system based on speech recognition. We focus on the task of automatically routing telephone calls based on user's fluently spoken response instead of touch tone menus in an interactive voice response system. Vector-based call routing algorithm is investigated and mapping table for key term is suggested. Korail database collected by KT is used for call routing experiment. We evaluate call-classification experiments for transcribed text from Korail database. In case of small training data, an average call routing error reduction rate of 14% is observed when mapping table is used.

• Journal of Chest Surgery
• /
• 제7권2호
• /
• pp.175-178
• /
• 1974

The patient was 21-year old male who had gastrostomy and tracheostomy after swallowing lye-stuff in July 1971. He could restore his normal voice and breathing after removal of his destructed epiglottis obstructing his upper airway two years later. Pharyngogastrostomy was performed in Nov 1973. The esophagus which was totally obliterated in its full length was removed and the stomach was brought high up to the level of pharynx where it was anastomosed to the posterior wall of pharynx. His postoperative course was temporarily complicated by aspiration of small food into trachea which could be completely relieved with training, and he is doing his normal life quite well on the follow-up.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1994년도 Proceedings of the Korea Automatic Control Conference, 9th (KACC) ; Taejeon, Korea; 17-20 Oct. 1994
• /
• pp.522-526
• /
• 1994

There are great needs for checking machine operation status precisely in the iron and steel plants. Rotating machines such as pumps, compressors, and motors are the most important objects in the plant maintenance. In this paper back-propagation neural network is utilized in diagnosing rotating machines. Like the finger print or the voice print of human, the abnormal vibrations due to axis misalignment, shaft bending, rotor unbalance, bolt loosening, and faults in gear and bearing have their own spectra. Like the pattern recognition technique, characteristic. feature vectors are obtained from the power spectra of vibration signals. Then we apply the characteristic feature vectors to a back propagation neural net for the weight training and pattern recognition.