KR19980074392A - How voice recognition device works - Google Patents

How voice recognition device works Download PDF

Info

Publication number
KR19980074392A
KR19980074392A KR1019970010184A KR19970010184A KR19980074392A KR 19980074392 A KR19980074392 A KR 19980074392A KR 1019970010184 A KR1019970010184 A KR 1019970010184A KR 19970010184 A KR19970010184 A KR 19970010184A KR 19980074392 A KR19980074392 A KR 19980074392A
Authority
KR
South Korea
Prior art keywords
voice
command
recognition device
holder
present
Prior art date
Application number
KR1019970010184A
Other languages
Korean (ko)
Other versions
KR100335189B1 (en
Inventor
박찬석
Original Assignee
김영귀
기아자동차 주식회사
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 김영귀, 기아자동차 주식회사 filed Critical 김영귀
Priority to KR1019970010184A priority Critical patent/KR100335189B1/en
Publication of KR19980074392A publication Critical patent/KR19980074392A/en
Application granted granted Critical
Publication of KR100335189B1 publication Critical patent/KR100335189B1/en

Links

Classifications

    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00Speech recognition
    • G10L15/22Procedures used during a speech recognition process, e.g. man-machine dialogue
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00Speech recognition
    • G10L15/06Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
    • G10L15/065Adaptation
    • G10L15/07Adaptation to the speaker

Landscapes

  • Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Artificial Intelligence (AREA)
  • Lock And Its Accessories (AREA)

Abstract

본 발명은 음성인식 장치의 작동방법에 관한 것으로서, 본 발명의 목적은 음성 명령권자의 신체적 상태나 시간에 따라 변화하는 특성을 개선하여 화자인식률을 향상시킨 음성인식 장치의 작동방법을 제공함에 있다.The present invention relates to a method of operating a voice recognition device, and an object of the present invention is to provide a method of operating a voice recognition device to improve the speaker recognition rate by improving the characteristics that change with the physical state or time of the voice command holder.

상기와 같은 목적을 실현하기 위한 본 발명은 특정의 음성 명령권자를 입력하기 위해 비밀번호를 입력하여 확인될 경우 음성의 특징을 추출하여 데이타베이스에 구축하고 음성 명령권자의 음성명령중 음성변화의 상태가 음성 명령권자일 확률 이상이면 변화된 음성의 특징을 데이타베이스에 추가로 구축하는 음성 인식장치의 작동방법으로써 허가된 명령권자만이 음성명령을 내릴 수 있도록 하며 명령권자의 신체변화에 따라 변화를 음성의 변화 자료를 추가하여 명령권자의 인식율을 향상시킨다.The present invention for realizing the above object is to extract the feature of the voice to establish a database when the password is entered to confirm the specific voice command holder, and the state of the voice change of the voice command owner voice command right It is a method of operating the voice recognition device that adds the changed voice characteristics to the database if it is more than the probability of being a child, so that only authorized command holders can give voice commands. Improve the recognition rate of command holders.

Description

음성인식 장치 작동방법How voice recognition device works

본 발명은 음성인식 장치의 작동방법에 관한 것으로서, 보다 상세하게는 음성 명령권자의 신체적 상태나 시간에 따라 변화하는 특성을 개선하여 화자인식률을 향상시킨 음성인식 장치의 작동방법에 관한 것이다.The present invention relates to a method of operating a voice recognition device, and more particularly, to a method of operating a voice recognition device to improve the speaker recognition rate by improving the characteristics that change with the physical state or time of the voice command authority.

요즈음은 전자장치의 발달로 인하여 자동차의 성능향상을 위해 많은 부분에서 전자장치의 제어에 의존하고 있다. 특히 엔진의 성능향상을 위해 여러 가지의 센서가 사용되어 최적의 엔진 효율을 발생시키고 있어 차량의 신뢰도 높아지고 있다. 또한 이러한 전자장치의 발달은 실제적인 엔진작동 뿐만아니라 운전자의 안전을 도모하기 위한 안전장치나 운전의 편의를 위한 갖가지 부가장치, 주행장치 등에 적용되고 있다.Nowadays, due to the development of electronic devices, in order to improve the performance of automobiles, many rely on the control of electronic devices. In particular, a variety of sensors are used to improve the performance of the engine to generate the optimum engine efficiency, increasing the reliability of the vehicle. In addition, the development of such electronic devices is applied to not only actual engine operation but also safety devices for driving driver safety, various additional devices for driving convenience, and driving devices.

부가장치 중에서 새롭고 흥미로운 것 중 하나는 음성합성과 음석인식으로 자동차와 대화를 하면서 차량을 제어할 수 있도록 하는 것이다.One of the new and interesting additions is the ability to control the car while talking to the car with voice synthesis and speech recognition.

음성 인식장치는 운전자가 운전하면서 시야를 돌리지도 않고 주행에 필요한 손과 발동작을 모두 수행하면서 부가적인 장치를 제어하도록 하게된다.The voice recognition device allows the driver to control the additional device while performing all the hand and foot movements required for driving without turning the field of view.

그런데 자동차를 사용하는 사람이 불특정 다수인이기 때문에 음성 인식장치도 불특정 다수인의 음성을 인식하여 음성명령을 수행하도록 설치하고 있다.However, since the number of people using a car is unspecified, the voice recognition device is also installed to recognize a voice of an unspecified number of people and perform a voice command.

그러나 불특정 다수인의 음성을 인식하여 음성명령을 수행할 경우에는 운전자가 아닌 다른 사람 즉, 승객이나 외부인도 음성으로 차량을 제어할 수 있어 사고 유발의 가능성이 있다는 문제점이 있다.However, when a voice command is performed by recognizing an unspecified number of people's voices, there is a problem in that a person other than the driver, that is, a passenger or an external person, may control the vehicle by voice, which may cause an accident.

한편, 음성 명령권자일 경우에도 장시간 차량의 운전으로 인한 피로와 운전자의 신체변화에 따른 음성 특성의 변화로 음성 명령권자의 인식률이 저하된다는 문제점 또한 발생한다.On the other hand, even in the case of the voice command authority, there is also a problem that the recognition rate of the voice command authority is lowered due to the fatigue caused by driving of the vehicle for a long time and the change of the voice characteristics according to the body change of the driver.

본 발명은 상기와 같은 문제점을 해결하기 위해 창작된 것으로서, 본 발명의 목적은 허가된 음성 명령권자만의 음성을 인식하도록 하며 음성 명령권자의 신체변화에 따른 음성의 변화에도 음성 명령권자 인식률을 향상시킨 음성 인식장치의 작동방법을 제공함에 있다.The present invention has been made to solve the above problems, and an object of the present invention is to recognize only the authorized voice command holder voice and to improve the voice command holder recognition rate even with the voice change caused by the voice change of the voice command holder. The present invention provides a method of operating a recognition device.

도1은 본 발명에 의한 음성인식 장치의 작동방법을 나타낸 흐름도이다.1 is a flowchart illustrating a method of operating a voice recognition device according to the present invention.

상기와 같은 목적을 실현하기 위한 본 발명은 특정의 음성 명령권자를 입력하기 위해 비밀번호를 입력하여 확인될 경우 음성의 특징을 추출하여 데이타베이스에 구축하고 음성 명령권자의 음성명령중 음성변화의 상태가 음성 명령권자일 확률 높으면 변화된 음성의 특징을 데이타베이스에 추가로 구축하는 음성 인식장치의 작동방법이다.The present invention for realizing the above object is to extract the feature of the voice to establish a database when the password is entered to confirm the specific voice command holder, and the state of the voice change of the voice command owner voice command right It is a method of operating a speech recognition device that constructs additional features of a changed voice in a database when the probability of being high is high.

상기와 같이 이루어진 본 발명의 작동방법을 상세히 설명하면 명령권자를 등록할 경우에는 음성 인식장치에 입력할 수 있는 비밀번호를 입력하여 일치할 경우에는 명령권자의 음성을 입력받아 명령권자의 음성에서 특징을 추출하여 테이타베이스를 구축한다. 그리고 음성명령에 의한 명령입력중 운전자의 신체변화에 따라 변하는 음성특징을 추출하여 명령권자일 확률이 높으면 계속 데이타베이스에 구축하여 명령권자의 신체변화에 대응하도록 하여 화자인식률을 향상시킬 수 있게 된다.Referring to the operation method of the present invention made in detail as described above, when registering the command holder, if the password is input by the voice recognition device to match the command owner's voice is input to extract the features from the command owner's voice Build your base. In addition, if the probability of being the command authority is high by extracting the voice feature that is changed according to the driver's body change during the command input by the voice command, it is possible to improve the speaker recognition rate by constructing the database to cope with the body change of the command authority.

이하, 본 발명의 바람직한 실시예를 첨부된 도면을 참조하여 설명한다. 또한 본 실시예는 본 발명의 권리범위를 한정하는 것은 아니고, 단지 예시로 제시된 것이다.Hereinafter, exemplary embodiments of the present invention will be described with reference to the accompanying drawings. In addition, this embodiment is not intended to limit the scope of the present invention, but is presented by way of example only.

도1은 본 발명 따른 음성 인식장치의 작동방법에 따라 실시된 흐름도이다. 도1에 도시된 바와 같이 명령권자를 초기 등록하거나 추가 등록을 할 겻인가를 판단한다(S10). 명령권자를 등록할 경우에는 등록모드(S20)로 리턴되어 등록이 허가된 사람인지 확인하기 위해 비밀번호를 입력받는다(S210). 입력된 비밀번호를 비교하여 허가된 비밀번호인지 판단한다(S220). 판단한 결과 허가되지 않은 사람일 경우에는 초기의 명령권자 등록여부를 묻는 단계(S10)로 리턴되지만 비밀번호가 일치하여 허가된 사람일 경우에는 명령권자의 음성을 입력받는다(S230). 입력받은 명령권자의 음성에서 다른사람과 다른 특징을 찾아낸다(S240). 찾아낸 음성특징은 데이타베이스를 형성하여 기억시킨다(S250). 그리고 추가적으로 차량을 사용할 수 있는 다른 명령권자의 음성을 등록할 것인가를 묻게된다(S260). 여기서 더이상의 추가 등록이 없을 경우에는 종료되며 명령권자를 추가등록 할 경우에는 다른 명령권자의 음성을 입력하는 단계(S230)로 리턴된다.1 is a flowchart performed according to a method of operating a speech recognition apparatus according to the present invention. As shown in FIG. 1, it is determined whether to initially register or additionally register the command holder (S10). When registering the command holder is returned to the registration mode (S20) and receives a password to check whether the registration is authorized (S210). It is determined whether the password is authorized by comparing the input password (S220). If it is determined that the person is not authorized, the user is returned to the initial step of asking whether to register the commander (S10), but if the password is the same and the authorized person receives the commander's voice (S230). In the voice of the command holder received the input and finds a different feature (S240). The found voice feature forms and stores a database (S250). In addition, it is asked whether to register the voice of another command authority who can use the vehicle (S260). Here, if there is no further registration, the process is terminated, and when additionally registering the command holder, the method returns to step S230 of inputting the voice of another command owner.

한편, 명령권자를 등록하지 않고 음성명령을 수행하고자 하면 음성 인식모드로 리턴된다(S30). 음성 인식모드에서는 명령을 수행하기 위해 등록된 음성 명령권자인지를 확인하기 위해 음성을 입력받는다(S310). 이 때 입력받은 음성중에서 특징을 찾아낸다(S320). 그리고 찾아낸 특징과 데이타베이스에 구축된 음성데이타의 특징을 비교하여 등록된 명령권자인지 판단한다(S330). 명령권자임을 판단하여 명령권자가 아닐경우에는 초기의 명령권자 등록을 할 것인지 묻는 단계(S10)로 리턴되며 명령권자일 경우에는 수행할 명령을 입력받는다(S340). 입력된 명령을 수행하기에 앞서 입력된 명령의 음성중 명령권자이기는 하지만 명령권자의 신체변화에 의해 변화된 음성일 경우 명령권자일 확률이 설정치 높은지 판단한다(S350). 이 때 명령권자일 확률은 높지만 음성의 특징에 변화가 있을 경우에는 데이타베이스에 명령권자의 음성변화를 추가로 업데이트한다(S360). 그런다음 입력된 음성명령을 분석하여 명령을 수행하고 종료된다(S370).On the other hand, if the user wants to perform a voice command without registering the command holder is returned to the voice recognition mode (S30). In the voice recognition mode, a voice is input to confirm whether the voice command holder is registered to perform the command (S310). At this time, the feature is found in the received voice (S320). Then, it is determined whether the registered command authority is compared by comparing the found feature with the feature of the voice data constructed in the database (S330). If it is determined that the command holder is not the command holder is returned to step S10 asking whether to register the initial command holder (S10), if the command holder receives a command to perform (S340). Before performing the input command, if the voice of the input command is the command right but the voice is changed by the body change of the command right, it is determined whether the probability of being the command right is high (S350). At this time, if there is a high probability of being the command authority, but there is a change in the characteristics of the voice, the voice change of the command authority is further updated in the database (S360). Then, the input voice command is analyzed and the command is terminated (S370).

상기 실시예는 성공적으로 수행했을 경우 종료되고 있으나 전원이 공급되는 이상 계속적으로 작동시키기 위해서 전원의 온오프를 비교하여 오프일 경우에만 종료하고 전원이 온일 경우에는 명령권자를 추가하기위해 묻는 단계(S260)와 음성명령을 수행하는 단계(S370) 이후 새로운 음성명령의 입력이나 명령권자를 추가를 위해 묻는 명령권자를 등록할 것인지 판단하는 단계(S10)로 리턴되도록 함이 바람직하다.The above embodiment is terminated when it is successfully performed, but in order to continue to operate as long as the power is supplied, comparing the power on and off ends only when the power is off and asks to add a command holder when the power is on (S260). After the step of performing a voice command (S370) it is preferable to return to step (S10) to determine whether to register a command holder asking for input of a new voice command or the command holder to add.

그리고 명령권자를 새로이 등록할 것인지 묻는 단계(S260)에서 초기설정치를 음성을 인식하는 단계(S30)로 리턴되도록 하고 필요에 의해 명령권자를 등록할 때만 스위치를 조작하여 등록모드(S20)로 리턴되도록 함이 바람직하다.Then, in step S260, when the command holder is newly registered, the initial set value is returned to the step of recognizing the voice, and if necessary, only when the command owner is registered, the switch is returned to the registration mode S20. desirable.

상기한 바와 같이 본 발명은 음성 인식장치에서 특정인의 허가된 명령권자만이 음성명령을 내릴 수 있도록 하며 명령권자의 신체변화에 따른 음성의 변화를 계속 추가하여 명령권자의 인식률을 증가시켜 타인의 음성으로 인한 오동작과 개인의 음성변화로 인한 오동작을 줄일 수 있다는 이점이 있다.As described above, the present invention allows only the authorized command holder of a specific person to issue a voice command in the voice recognition device and increases the recognition rate of the command holder by continuously adding a voice change according to the body change of the command holder, thereby causing malfunction of the voice of another person. There is an advantage that can reduce the malfunction caused by the change of voice of individuals and individuals.

Claims (1)

특정의 음성 명령권자를 입력하기 위해 비밀번호를 입력하여 확인될 경우 음성의 특징을 추출하여 데이타베이스에 구축하고 음성 명령권자의 음성명령중 음성변화의 상태가 음성 명령권자일 확률이 높으면 변화된 음성의 특징을 데이타베이스에 추가로 구축하는 음성 인식장치의 작동방법.If it is confirmed by inputting a password to input a specific voice command holder, the voice feature is extracted and built into the database. How to operate the speech recognition device in addition to building.
KR1019970010184A 1997-03-25 1997-03-25 Method of operating voice recognition unit KR100335189B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1019970010184A KR100335189B1 (en) 1997-03-25 1997-03-25 Method of operating voice recognition unit

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1019970010184A KR100335189B1 (en) 1997-03-25 1997-03-25 Method of operating voice recognition unit

Publications (2)

Publication Number Publication Date
KR19980074392A true KR19980074392A (en) 1998-11-05
KR100335189B1 KR100335189B1 (en) 2002-06-20

Family

ID=37479749

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1019970010184A KR100335189B1 (en) 1997-03-25 1997-03-25 Method of operating voice recognition unit

Country Status (1)

Country Link
KR (1) KR100335189B1 (en)

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5539861A (en) * 1993-12-22 1996-07-23 At&T Corp. Speech recognition using bio-signals
KR0169592B1 (en) * 1996-09-05 1999-03-20 양승택 Performance enhancing method for voice recognition device using adaption of voice characteristics
JPH10282988A (en) * 1997-04-08 1998-10-23 Kenwood Corp Speech recognition device

Also Published As

Publication number Publication date
KR100335189B1 (en) 2002-06-20

Similar Documents

Publication Publication Date Title
US10604111B2 (en) Vehicle control apparatus, vehicle control method, and recording medium storing program
EP1082671B1 (en) Handwritten and voice control of vehicle appliance
US5774858A (en) Speech analysis method of protecting a vehicle from unauthorized accessing and controlling
US7158871B1 (en) Handwritten and voice control of vehicle components
US20060173680A1 (en) Partial spelling in speech recognition
US9881609B2 (en) Gesture-based cues for an automatic speech recognition system
JPH11231886A (en) Registered name recognition device
CN105222797B (en) Utilize the system and method for oral instruction and the navigation system of partial match search
JP2005128307A (en) Vocal print authentication system
JP2003202897A (en) Speech recognizing device for on-vehicle equipment
JP2008501992A (en) Dialog system and method for user authentication
CN105103087B (en) It is a kind of to be used to showing and/or exporting the method for help and be allocated to the device of vehicle
US11508370B2 (en) On-board agent system, on-board agent system control method, and storage medium
CN113270093A (en) Proxy device, proxy system, and non-transitory recording medium
JP2016133378A (en) Car navigation device
JP7347244B2 (en) Agent devices, agent systems and programs
KR100335189B1 (en) Method of operating voice recognition unit
JP2005003997A (en) Device and method for speech recognition, and vehicle
CN106850673A (en) User vehicle accessing method and device based on Application on Voiceprint Recognition
JP2003228395A (en) Method and apparatus for equipment control by voice recognition
JP2000122685A (en) Navigation system
JP2007034198A (en) Speech recognition system and mobile terminal device used therefor
CN110111781B (en) Information processing apparatus, information processing method, and computer program
JP2006023444A (en) Speech dialog system
JP2005114964A (en) Method and processor for speech recognition

Legal Events

Date Code Title Description
A201 Request for examination
E701 Decision to grant or registration of patent right
GRNT Written decision to grant
LAPS Lapse due to unpaid annual fee