JPH0277799A - On-vehicle speech recognition device - Google Patents
On-vehicle speech recognition deviceInfo
- Publication number
- JPH0277799A JPH0277799A JP63228842A JP22884288A JPH0277799A JP H0277799 A JPH0277799 A JP H0277799A JP 63228842 A JP63228842 A JP 63228842A JP 22884288 A JP22884288 A JP 22884288A JP H0277799 A JPH0277799 A JP H0277799A
- Authority
- JP
- Japan
- Prior art keywords
- microphone
- driver
- signal
- voice
- vehicle
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 238000000605 extraction Methods 0.000 claims description 13
- 230000005236 sound signal Effects 0.000 claims description 9
- 230000004044 response Effects 0.000 claims description 2
- 230000008030 elimination Effects 0.000 abstract 2
- 238000003379 elimination reaction Methods 0.000 abstract 2
- 238000010586 diagram Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000000034 method Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Abstract
Description
【発明の詳細な説明】
[産業上の利用分野コ
本発明は音声によって車輌の諸機器を制御するだめの音
声認識装置に関する。DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a voice recognition device for controlling various devices of a vehicle by voice.
[発明の概要]
音声信号を入力とする音声入力部と、その音声信号の特
徴抽出を行なう特徴抽出部と、その特徴抽出部で得られ
た入カバターンメモリと登録パターンメモリの類似度計
算を行なう認識処理部を持ち、その計算結果に応じた制
御出力を発生する。[Summary of the invention] An audio input section that receives an audio signal as input, a feature extraction section that extracts features of the audio signal, and similarity calculation between an input pattern memory and a registered pattern memory obtained by the feature extraction section. It has a recognition processing unit that performs the calculation, and generates a control output according to the calculation result.
制御のための音声以外の音源を雑音源と見なし、運転者
以外の話者用マイクロフォン、車の騒音用マイクロフォ
ンとして制御のための音声入力用マイクロフォン以外の
マイクロフォンも設置される。Sound sources other than voice for control are regarded as noise sources, and microphones other than the voice input microphone for control are also installed as microphones for speakers other than the driver and microphones for vehicle noise.
音声入力部は、複数個のマイクロフォンに対応して複数
個の増幅器と AGC回路を持ち、それらのA、 G
C回路の出力を入力とする雑音消去部が備えられる。音
響機器出力信号は消去部の入力群の一つである。The audio input section has multiple amplifiers and AGC circuits corresponding to multiple microphones, and their A, G
A noise canceling section is provided which receives the output of the C circuit as an input. The audio equipment output signal is one of the input groups of the canceling unit.
[従来の技術]
最近の半導体技術の進歩には目覚ましいものが有り、音
声認識装置においても、認識語数が限定され、しかも話
者が特定されている場合には、比較的小型で低価格のも
のが実用化されている。音声により機器を制御する場合
、音声を入力する方法は非常に重要な問題である。人間
は多少の距離があっても、多少の騒音下でも音声を聞き
分けることができる。 S/N比の悪い音声に対する電
子回路の認識能力は現状では非常に低い。これを改善す
るために次の二つの手段が採られている。[Prior art] Recent advances in semiconductor technology have been remarkable, and even in speech recognition devices, when the number of words to be recognized is limited and the speaker is specified, relatively small and low-cost devices are available. has been put into practical use. When controlling equipment by voice, the method of inputting voice is a very important issue. Humans can distinguish voices even from a certain distance and in some noise. At present, the ability of electronic circuits to recognize speech with a poor S/N ratio is extremely low. In order to improve this, the following two measures are taken.
(1)話者の近傍にマイクロフォンを置き、音声信号を
被制御装置内に有る認識装置に伝送する。(1) A microphone is placed near the speaker and the audio signal is transmitted to a recognition device in the controlled device.
(2)話者の近傍にマイクロフォンを置き、そのマイク
ロフォンに組み込まれた認識装置から制御信号が被制御
装置へ伝送される。(2) A microphone is placed near the speaker, and a control signal is transmitted from a recognition device built into the microphone to a controlled device.
すなわち、(1)、(2)とも話者との物理的距離を小
さくすることによって、 S/N比を上げることを狙っ
ている。これらの方式は、例えばTV等の民生機器につ
いて、話者の手もとで行ない得るので、マイクロフォン
と話者の距離が一定でしかも短距離になり、高SlN比
が得られる。That is, both (1) and (2) aim to increase the S/N ratio by reducing the physical distance to the speaker. These methods can be performed at the hands of the speaker in consumer equipment such as TVs, so the distance between the microphone and the speaker is constant and short, and a high SIN ratio can be obtained.
現在実用化されている音声認識装置は単語を単位とした
パターンマツチングを行なう話者限定の単語音声認識シ
ステムである。Speech recognition devices currently in practical use are speaker-specific word speech recognition systems that perform pattern matching on a word-by-word basis.
第4図は音声認識装置の一般的なブロック図で、図中、
21 は音声入力、22 は音声入力部、23 は特徴
抽出部、24 は認識処理部、25は入カバターンメモ
リ、26 は登録パターンメモリ、27 は制御出力を
表わす。Figure 4 is a general block diagram of a speech recognition device.
21 is a voice input, 22 is a voice input section, 23 is a feature extraction section, 24 is a recognition processing section, 25 is an input cover pattern memory, 26 is a registered pattern memory, and 27 is a control output.
このシステムには、登録モードと認識モードを持ち、登
録モードは認識しようとする単語音声を予め登録するも
ので、登録音声から特徴抽出部23 により抽出された
音声の特徴が時系列的パターンとして登録パターンメモ
リ 26 にファイルされる。したがって、特徴抽出部
23 にはサンプリング部と A/D変換部が含まれる
。This system has a registration mode and a recognition mode. In the registration mode, the word sounds to be recognized are registered in advance, and the features of the sounds extracted by the feature extraction unit 23 from the registered sounds are registered as time-series patterns. Filed in pattern memory 26. Therefore, the feature extraction section 23 includes a sampling section and an A/D conversion section.
認識モードにおいては入力音声から同様に抽出された音
声の特徴パターンが入力パターンメモリ25 に記憶さ
れた後、登録パターンと入カバターンの類似度が認識処
理部24 で計算される。In the recognition mode, after the voice feature pattern similarly extracted from the input voice is stored in the input pattern memory 25, the similarity between the registered pattern and the input pattern is calculated by the recognition processing section 24.
最も類似度の高い登録パターンが入力音声と一致するも
のとして、特定すなわち認識され、これに対応する出力
がなされる。このシステムでは、特徴抽出部23 に至
るまでの音声入力の S/N比が、正確な特徴抽出動作
に非常に大きく影響を及ぼし、ひいては装置全体の認識
率にも大きく影響を与える。The registered pattern with the highest degree of similarity is identified or recognized as matching the input voice, and a corresponding output is generated. In this system, the S/N ratio of the voice input up to the feature extractor 23 has a very large effect on accurate feature extraction operations, and in turn, has a large effect on the recognition rate of the entire device.
[発明が解決しようとする課題]
しかしながら、音声認識装置を車載用システムで用いる
場合を考えると1話者すなわち運転者とマイクロフォン
の位置関係は限られ、しかもマイクロフォンと話者の発
音源の角度はある範囲内で常に変化し、一定でない。さ
らに車内外騒音が有り、高い S/N比を持つ音声入力
信号を得ることは困難である。また運転者にマイクロフ
ォンを装着することは一層困難である。[Problems to be Solved by the Invention] However, when considering the case where a speech recognition device is used in an in-vehicle system, the positional relationship between one speaker, that is, the driver, and the microphone is limited, and the angle between the microphone and the speaker's sound source is limited. It constantly changes within a certain range and is not constant. Furthermore, there is noise inside and outside the vehicle, making it difficult to obtain an audio input signal with a high S/N ratio. It is also more difficult to equip the driver with a microphone.
[発明の目的コ
本発明の目的は、前記従来技術の欠点を無くし、車内で
も高S/N比を持ち、高い認識率を持つ車載用音声認識
装置を提供することである。[Object of the Invention] An object of the present invention is to provide an in-vehicle speech recognition device that eliminates the drawbacks of the prior art and has a high S/N ratio and a high recognition rate even in a car.
[課題を解決するための手段]
上記目的を達成するために、本発明による車載用音声認
識装置は、音声信号集音手段と、周囲雑音集音手段と、
上記音声信号集音手段から得られた信号と上記周囲雑音
集音手段から得られた信号から音声信号以外を消去する
雑音消去手段と、上記雑音消去手段を介した信号の特徴
抽出を行なう特徴抽出手段と、該特徴抽出手段で得られ
たパターンと登録パターンの類似比較を行なう認識処理
部と、上記類似比較に対応して制御出力を発生する手段
とを含むことを要旨とする。[Means for Solving the Problems] In order to achieve the above object, an in-vehicle speech recognition device according to the present invention includes a voice signal collection means, an ambient noise collection means,
noise erasing means for erasing non-audio signals from the signal obtained from the audio signal collecting means and the signal obtained from the ambient noise collecting means; and feature extraction for extracting features of the signal via the noise erasing means. A recognition processing unit that performs a similarity comparison between a pattern obtained by the feature extraction means and a registered pattern, and a means that generates a control output in response to the similarity comparison.
[作用]
話者用マイクロフォンが運転者からある程度前れていて
、しかも話者の発音源との角度が一定でなく、話者用マ
イクロフォンから集音された音声信号そのものの S/
N比は低く、このま\では認識処理できない。しかし、
運転者以外の音声および騒音をそれぞれ専用のマイクロ
フォンにより得られた雑音信号と音響出力信号を雑音消
去部に入力し、運転者用マイクロフォンによって得られ
た信号に混入している雑音を取り除くことによって」―
紀行識処理を可能とする。[Effect] If the speaker's microphone is located in front of the driver to some extent, and the angle with the speaker's sound source is not constant, the S/S of the voice signal itself collected from the speaker's microphone may vary.
The N ratio is low, and recognition processing is not possible at this rate. but,
By inputting the noise signal and acoustic output signal obtained from voices and noises other than the driver's voice and noise through dedicated microphones to a noise canceling section, and removing the noise mixed in the signal obtained by the driver's microphone. ―
Enables travel knowledge processing.
「実施例コ
以下に、図面を参照しながら、実施例を用いて本発明を
一層詳細に説明するが、それらは例示に過ぎず、本発明
の枠を越えることなしにいろいろな変形や改良があり得
ることは勿論である。``Example'' The present invention will be explained in more detail below using examples with reference to the drawings, but these are merely illustrative and various modifications and improvements can be made without going beyond the scope of the present invention. Of course it is possible.
第1図は本発明による車載用音声認識装置の音声入力部
のブロック図で、図中、 」 は話者(運転者)用マイ
クロフォン、 2は運転者以外の話者用マイクロフォン
、 3 は車輌の騒音用マイクロフォン、 4 は増幅
器と AGC回路、5 は音響機器出力(スピーカ出力
)信号、 6は雑音消去部、7 は増幅器、8 は音声
入力部、9 は特徴抽出部へ至る端子を表わす。FIG. 1 is a block diagram of the voice input section of the in-vehicle voice recognition device according to the present invention. In the figure, "" is the microphone for the speaker (driver), 2 is the microphone for speakers other than the driver, and 3 is the microphone for the vehicle. A noise microphone, 4 an amplifier and an AGC circuit, 5 an audio equipment output (speaker output) signal, 6 a noise canceling section, 7 an amplifier, 8 an audio input section, and 9 a terminal leading to a feature extraction section.
第2図は家庭用テレビジョン装置の音声入力部で、図中
、 11 は話者用マイクロフォン、12は増幅器、1
3 はAGC回路、14 は特徴抽出部へ至る端子、1
5 は音声入力部を示す。Figure 2 shows the audio input section of a home television device, in which 11 is a speaker microphone, 12 is an amplifier, 1
3 is the AGC circuit, 14 is the terminal leading to the feature extraction section, 1
5 indicates an audio input section.
第3図は本発明による音声認識装置のマイクロフォン配
置例を示す平面図で、図中、31 は車載用音声認識応
用装置(運転者マイクロフォン35 内臓)、32 は
車の騒音用マイクロフォン、33 は運転者以外の話者
用マイクロフォン、34 は自動車である。FIG. 3 is a plan view showing an example of the microphone arrangement of the voice recognition device according to the present invention. In the figure, 31 is an in-vehicle voice recognition application device (driver microphone 35 built-in), 32 is a car noise microphone, and 33 is a driving device. The microphone for speakers other than the speaker, 34, is a car.
以下上記実施例の動作を説明する。The operation of the above embodiment will be explained below.
第1図において、制御のための音声入力用マイクロフォ
ン 1 が運転者用として設置され、さらに運転者以外
の話者用マイクロフォン 2 、車輌の騒音用マイクロ
フォン 3 が追加設置されている。また、それぞれの
マイクロフォンで得られた信号の他に音響機器出力信号
が直接雑音消去部6 の入力となっている。In FIG. 1, a voice input microphone 1 for control is installed for the driver, and a microphone 2 for speakers other than the driver and a microphone 3 for vehicle noise are additionally installed. In addition to the signals obtained by the respective microphones, audio equipment output signals are directly input to the noise canceling section 6.
運転者用マイクロフォンを車輌のハンドル部に設置し、
そのマイクロフォンに内蔵されている送信機から音響機
器に内蔵されている受信機に送信するようにしてもよい
。この方式によると、運転者用マイクロフォンをハンド
ル部に設置することで、運転者までの距離を小さくし、
一定にすることができ、より高い S/N比が得ら九る
。Install the driver's microphone on the steering wheel of the vehicle.
The signal may be transmitted from a transmitter built into the microphone to a receiver built into the audio equipment. According to this method, by installing the driver's microphone on the steering wheel, the distance to the driver can be reduced.
can be kept constant, resulting in a higher S/N ratio.
[発明の効果]
以」―説明した通り、本発明によれば、運転者具−7=
外の音声や騒音の環境のもとにおいても、S/N比の高
い音声信号が得られ、音声認識装置の認識率の向上に大
きく寄与できる。[Effects of the Invention] As explained above, according to the present invention, an audio signal with a high S/N ratio can be obtained even in an environment with external sounds and noise, and the audio signal can be improved. This can greatly contribute to improving the recognition rate of the recognition device.
第1図は本発明による車載用音声認識装置の音声入力部
のブロック図、第2図は家庭用テレビジョン装置の音声
入力部ブロック図、第3図は本発明による音声認識装置
のマイクロフォン配置例を示す平面図、第4図は音声認
識装置の一般的なブロック図である。
1・・・・・・・運転者用マイクロフォン、2・・・・
・運転者以外の話者用マイクロフォン、3・・・・・
車輌の騒音用マイクロフォン、4・・・・・・・・増
幅器とAGC回路、5・・・・・・・音響機器出力(ス
ピーカ出力)信号、6・・・・・・・・雑音消去部、7
・・・・・・・増幅器、8・・・・・・・・音声入力部
、9・・・・・特徴抽出部へ至る端子。
特許出願人 クラリオン株式会社FIG. 1 is a block diagram of the voice input section of the in-vehicle voice recognition device according to the present invention, FIG. 2 is a block diagram of the voice input section of a home television device, and FIG. 3 is an example of microphone arrangement of the voice recognition device according to the present invention. FIG. 4 is a general block diagram of a speech recognition device. 1... Driver microphone, 2...
・Microphone for speakers other than the driver, 3...
Vehicle noise microphone, 4...Amplifier and AGC circuit, 5...Audio equipment output (speaker output) signal, 6...Noise canceller, 7
......Amplifier, 8...Audio input section, 9...Terminal leading to feature extraction section. Patent applicant Clarion Co., Ltd.
Claims (1)
囲雑音集音手段から得られた信号から音声信号以外を消
去する雑音消去手段、 (d)上記雑音消去手段を介した信号の特徴抽出を行な
う特徴抽出手段、 (e)該特徴抽出手段で得られたパターンと登録パター
ンの類似比較を行なう認識処理部、および (f)上記類似比較に対応して制御出力を発生する手段 を含むことを特徴とする車載用音声認識装置。[Claims] (a) Audio signal collection means; (b) Ambient noise collection means; (c) Signal obtained from the audio signal collection means and signal obtained from the ambient noise collection means. (d) feature extraction means for extracting features of the signal via the noise cancellation means; (e) similarity comparison between the pattern obtained by the feature extraction means and the registered pattern; and (f) means for generating a control output in response to the similarity comparison.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP63228842A JPH0277799A (en) | 1988-09-14 | 1988-09-14 | On-vehicle speech recognition device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP63228842A JPH0277799A (en) | 1988-09-14 | 1988-09-14 | On-vehicle speech recognition device |
Publications (1)
Publication Number | Publication Date |
---|---|
JPH0277799A true JPH0277799A (en) | 1990-03-16 |
Family
ID=16882719
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
JP63228842A Pending JPH0277799A (en) | 1988-09-14 | 1988-09-14 | On-vehicle speech recognition device |
Country Status (1)
Country | Link |
---|---|
JP (1) | JPH0277799A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0472356A1 (en) * | 1990-08-16 | 1992-02-26 | Fujitsu Ten Limited | Speech recognition apparatus for a vehicle, using a microphone arrangement to determine the seat from which a command is generated |
US7016836B1 (en) * | 1999-08-31 | 2006-03-21 | Pioneer Corporation | Control using multiple speech receptors in an in-vehicle speech recognition system |
-
1988
- 1988-09-14 JP JP63228842A patent/JPH0277799A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0472356A1 (en) * | 1990-08-16 | 1992-02-26 | Fujitsu Ten Limited | Speech recognition apparatus for a vehicle, using a microphone arrangement to determine the seat from which a command is generated |
US7016836B1 (en) * | 1999-08-31 | 2006-03-21 | Pioneer Corporation | Control using multiple speech receptors in an in-vehicle speech recognition system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9881616B2 (en) | Method and systems having improved speech recognition | |
WO2020173156A1 (en) | Method, device and electronic device for controlling audio playback of multiple loudspeakers | |
CN1288223A (en) | Device adaptive for direction characteristic used for speech voice control | |
DE69531861D1 (en) | VOICE-CONTROLLED VEHICLE ALARM SYSTEM | |
JP2023159381A (en) | Sound recognition audio system and method thereof | |
CN101751918A (en) | Novel silencer and noise reduction method | |
JP2007180896A (en) | Voice signal processor and voice signal processing method | |
KR101704510B1 (en) | Method for enhancement of speech of interest, an apparatus for enhancement of speech of interest and a vehicle equipped with the apparatus | |
JP2019184809A (en) | Voice recognition device and voice recognition method | |
JPH1152976A (en) | Voice recognition device | |
CN215815174U (en) | Voice interaction system inside and outside vehicle and vehicle | |
JP2000148184A (en) | Speech recognizing device | |
JP3257832B2 (en) | Noise reduction circuit for speech recognition device | |
JP3411648B2 (en) | Automotive audio equipment | |
CN113053402A (en) | Voice processing method and device and vehicle | |
JP3774580B2 (en) | Voice input device | |
JPH0277799A (en) | On-vehicle speech recognition device | |
JP5377442B2 (en) | System that separates speech from noise by reference information | |
Okuno et al. | Robot audition: Missing feature theory approach and active audition | |
JP2000322074A (en) | Voice input section determination device, aural data extraction device, speech recognition device, vehicle navigation device and input microphone | |
JPH0396999A (en) | Sound collecting device | |
JPS5830800A (en) | Voice recognition system for vehicle | |
EP1091347A2 (en) | Multi-stage speech recognition | |
CN115762518A (en) | Vehicle external voice interaction system and method and vehicle applying same | |
WO2023104215A1 (en) | Methods for synthesis-based clear hearing under noisy conditions |