Abstract:
PURPOSE: A computer input and output control method and an apparatus thereof are provided to minimize restriction in a hardware module size and to maximize flexible management of various requirements, thereby reducing design costs. CONSTITUTION: A computer input and output control method applying VPX(Versatile Performance Switching) bus standard and FPGA(Field Programmable Gate Array) IP(Intellectual Property) core technology includes following steps. A single board computer receives power through a backplane board(S600). The single board computer checks whether the single board computer communicates with a different module or not through an input and output module(S610). The single board computer communicates with the different module(S620). The backplane board includes a bus based on VPX standard for high-speed data processing. The input and output module loads an IP core in an FPGA chip. [Reference numerals] (AA) Start; (BB) End; (S600) Step of receiving power through a backplane board; (S610) Step of checking whether communicating with a different module or not; (S620) Step of communicating with the different module after checking the communication with the different module through an input and output module
Abstract:
본 발명의 전투기 조종사 음성명령실행방법에서는 발화된 조종사의 음성명령어(전투기 명령어)가 단어단위별로 자동분할되어 단구간신호의 에너지 처리와 영 교차율(Zero Crossing Rate)처리로 묵음이 가려지고, MFCC(Mel-Frequency Cepstrum Coefficent)로부터 음성특성 파라미터(Parameter)가 결정됨으로써 각 단어별 인식 후보군이 설정되며, 설정된 단어별 인식 후보군이 구문트리(Verbal Tree)가 기반된 연결단어 정합기술로 연결단어 기반의 명령어로 이루어짐으로써 단어조합 수량을 크게 줄이고 명령어 인식률이 크게 높아지며, 특히 유사도 측정 횟수를 줄여줌으로써 단어 인식시간이 단축되는 특징을 갖는다.
Abstract:
According to a fighter pilot voice command execution method of the present invention, a voice command (fighter command) of a pilot is automatically divided into the word unit and a silent syllable is hidden with the energy processing of a short section signal and zero crossing rage processing. A voice characteristic parameter is determined from a mel-frequency cepstrum coefficient (MFCC) and a recognizing candidate group by the word is set. The recognizing candidate group by the word is formed of a command of a connection word base with a connection word matching technique based on a verbal tree. Therefore, word matching times are reduced dramatically and a recognizing rate of the command is highly increased. Especially, a word recognizing time is reduced by reducing the similarity measuring times.