CN110797009A - Aircraft cabin instruction recognition device to bakistan accent english - Google Patents
Aircraft cabin instruction recognition device to bakistan accent english Download PDFInfo
- Publication number
- CN110797009A CN110797009A CN201810781122.7A CN201810781122A CN110797009A CN 110797009 A CN110797009 A CN 110797009A CN 201810781122 A CN201810781122 A CN 201810781122A CN 110797009 A CN110797009 A CN 110797009A
- Authority
- CN
- China
- Prior art keywords
- english
- instruction
- module
- english instruction
- recognition
- 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
- 230000005236 sound signal Effects 0.000 claims abstract description 33
- 238000013500 data storage Methods 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims description 8
- 230000003993 interaction Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 241000282414 Homo sapiens Species 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/02—Feature extraction for speech recognition; Selection of recognition unit
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
- G10L2015/0635—Training updating or merging of old and new templates; Mean values; Weighting
- G10L2015/0636—Threshold criteria for the updating
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)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Toys (AREA)
Abstract
The invention discloses an airplane cockpit instruction recognition device aiming at Bakistan accent English, which mainly comprises a voice acquisition module, an English instruction recognition module, an input/output module, an audio data storage module and a model updating module. The audio data storage module is used for receiving the English instruction audio signal from the voice acquisition module and the English instruction recognition result from the English instruction recognition module, and in a non-model updating state, the English instruction audio signal and the English instruction recognition result are judged to be stored or discarded by combining the recognition confidence coefficient so as to generate data to be updated. And the model updating module receives the data to be updated from the audio data storage module in a model updating state and updates a matching model used by an English instruction recognition engine built in the English instruction recognition module according to the data to be updated. The invention has the advantages that: the accuracy of the Bakistan accent English under the universal English model and the recognition engine is improved.
Description
Technical Field
The invention relates to an airplane cockpit instruction recognition device aiming at a Basstein accent English language.
Background
The voice interaction is the most natural interaction mode of human beings, so the voice input mode is also applied to the human-computer interaction process, and particularly under the condition that key input or touch screen input is inconvenient, the voice input can greatly improve the input efficiency. The technical background is that the voice control mode instead of the manual mode is adopted to complete part of the cabin operation in the aircraft cabin, which is a natural functional requirement upgrade. The airplane cabin command recognition device is internally provided with a voice recognition engine and a model and can be used for recognizing the collected input voice command. And when the voice command needing to be recognized is an English command, the corresponding model is an English model. Similar to Chinese speech in dialects of various regions, English in different regions also has own accent.
Disclosure of Invention
The invention aims to solve the problem that the accuracy of the Basstein accent English is low under a general English model and a recognition engine aiming at the position with obvious pronunciation difference between the Basstein accent English and standard English (American English), and provides a novel airplane cabin instruction recognition device aiming at the Basstein accent English.
In order to achieve the purpose, the technical scheme of the invention is as follows: an airplane cockpit instruction recognition device aiming at the Bakistan accent English comprises,
the voice acquisition module is used for picking up English instruction voice provided by the outside and generating an English instruction audio signal corresponding to the English instruction voice, and the English instruction language comprises Pakistan accent English;
the English instruction recognition module is used for receiving the English instruction audio signal from the voice acquisition module and matching a corresponding English instruction recognition result in a built-in English instruction recognition engine according to the English instruction audio signal;
the input and output module is used for receiving the English instruction recognition result from the English instruction recognition module in a non-model updating state, and matching the corresponding instruction code in a built-in instruction code query engine according to the English instruction recognition result so as to output an external instruction;
the audio data storage module is used for receiving the English instruction audio signal from the voice acquisition module and the English instruction recognition result from the English instruction recognition module, and making a judgment of saving or discarding the English instruction audio signal and the English instruction recognition result in a non-model updating state by combining recognition confidence to generate data to be updated; and the number of the first and second groups,
and the model updating module is used for receiving the data to be updated from the audio data storage module in a model updating state and updating a matching model used by an English instruction recognition engine built in the English instruction recognition module according to the data to be updated.
As a preferable scheme of the airplane cabin command recognition device aiming at the bakistan accent English, the airplane cabin command recognition device further comprises,
and the keyboard mouse and screen display module is used for selectively switching the instruction recognition device to a model updating state or a non-model updating state.
As a preferable scheme of the airplane cabin command recognition device aiming at the bakistan accent English, the airplane cabin command recognition device further comprises,
and the power supply module is used for supplying power to the voice acquisition module, the English instruction identification module, the instruction input and output functional module, the audio data storage module, the model updating module and the keyboard, mouse and screen display module.
The invention also provides an airplane cockpit instruction identification method aiming at the Bakistan accent English, which comprises the following steps,
step S1, providing the command recognition device;
step S2, collecting English instruction voice provided by the outside by the voice collecting module and generating English instruction audio signal corresponding to the English instruction voice;
step S3, receiving the English instruction audio signal from the voice acquisition module by an English instruction recognition module, and matching a corresponding English instruction recognition result in a built-in English instruction recognition engine according to the English instruction audio signal; and the number of the first and second groups,
step S4, the input/output module receives the english instruction recognition result from the english instruction recognition module and matches the corresponding instruction code in the built-in instruction code query engine according to the english instruction recognition result to output the external instruction.
The invention also provides an airplane cockpit instruction identification method aiming at the Bakistan accent English, which comprises the following steps,
step S1, providing the command recognition device;
step S2, collecting English instruction voice provided by the outside by the voice collecting module and generating English instruction audio signal corresponding to the English instruction voice;
step S3, the English instruction recognition module receives the English instruction audio signal from the voice acquisition module and matches the English instruction recognition result corresponding to the English instruction audio signal in a built-in English instruction recognition engine according to the English instruction audio signal;
step S4, the audio data storage module receives the English instruction audio signal from the voice acquisition module and the English instruction recognition result from the English instruction recognition module, and in a model updating state, the English instruction audio signal and the English instruction recognition result are judged to be stored or discarded according to recognition confidence, so that data to be updated are generated; and the number of the first and second groups,
step S5, the model updating module receives the data to be updated from the audio data storage module in the model updating state and updates the matching model used by the english instruction recognition engine built in the english instruction recognition module according to the data to be updated.
Compared with the prior art, the invention has the advantages that: the accuracy of the Bakistan accent English under the universal English model and the recognition engine is improved.
Drawings
Fig. 1 is a schematic structural diagram according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail below with reference to specific embodiments and drawings.
Referring to fig. 1, an aircraft cabin command recognition apparatus for the bakistan accent english is shown. The method comprises the following steps: the device comprises a voice acquisition module 1, an English instruction recognition module 2, a model updating module 3, a data storage module 4, an instruction input/output module 5, a keyboard, mouse and screen display module 6, a device control switch 7 and a power supply module 8.
The complete instruction recognition and model update process is described as follows:
1. the whole airplane cabin English instruction recognition device is started through the control switch 7, and the power supply module 8 starts to supply power to the recognition device.
2. The user speaks a specific English instruction, the voice acquisition module 1 carries out pickup, and the processed audio signal is sent to the English instruction recognition module 2 and the data storage module 4.
3. The English instruction identification module 2 identifies the input audio signal and sends the identification result to the instruction input and output module 5 and the data storage module 4.
4. And the input and output module 5 matches the corresponding instruction code according to the identification result and outputs the instruction code to the outside.
5. And the data storage module 4 judges whether the received audio data and the recognition result are stored or discarded according to the confidence coefficient of the recognition result. And if the model is stored, storing the two corresponding data as the marking data of the model updating.
6. In the case of accessing the keyboard and mouse and the screen display module 6, the model update state can be entered by keyboard and mouse operation. Under the state of updating the model, the user can repeatedly and continuously carry out recording operation according to the prompt of the screen display interface. All the recording operations are subjected to three steps of 2, 3 and 5, and effective data are stored in the data storage module 4;
7. when the saved effective data reaches a certain amount, the model updating state can be entered through keyboard and mouse operation. In the model updating state, the user updates the English model used in the English instruction recognition module 2 through a model updating button on the screen display interface. Note that this step operation and the english instruction recognition step cannot be performed simultaneously, and the user is asked to schedule time for updating.
8. The whole airplane cabin English instruction recognition device is turned off through the control switch 7, and the power supply module 8 stops supplying power to the recognition device.
The confidence coefficient is a type of scoring mechanism of a built-in English instruction recognition engine in the matching process, (the scoring range can be normalized to an interval of 0-100 points), and the score size can be used for judging the reliability degree of the matching result. By setting a threshold (for example, selecting 90 as the threshold), selecting the english instruction audio signal and the english instruction recognition result whose confidence coefficient of the matching result is higher than the threshold to store, and discarding the one whose confidence coefficient is lower than the threshold, the strategy can make the stored data have better quality and be suitable for being used as the update data. Since the users of the device are based on the Bakistan users, the standard accent English data of a small number of non-Bakistan users does not affect the effect of the subsequent model update module.
The foregoing merely represents embodiments of the present invention, which are described in some detail and detail, and therefore should not be construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (5)
1. An airplane cockpit command recognition device aiming at Backistan accent English is characterized by comprising,
the voice acquisition module is used for picking up English instruction voice provided by the outside and generating an English instruction audio signal corresponding to the English instruction voice, and the English instruction language is Pakistan accent English;
the English instruction recognition module is used for receiving the English instruction audio signal from the voice acquisition module and matching a corresponding English instruction recognition result in a built-in English instruction recognition engine according to the English instruction audio signal;
the input and output module is used for receiving the English instruction recognition result from the English instruction recognition module in a non-model updating state, and matching the corresponding instruction code in a built-in instruction code query engine according to the English instruction recognition result so as to output an external instruction;
the audio data storage module is used for receiving the English instruction audio signal from the voice acquisition module and the English instruction recognition result from the English instruction recognition module, and making a judgment of saving or discarding the English instruction audio signal and the English instruction recognition result in a non-model updating state by combining recognition confidence to generate data to be updated; and the number of the first and second groups,
and the model updating module is used for receiving the data to be updated from the audio data storage module in a model updating state and updating a matching model used by an English instruction recognition engine built in the English instruction recognition module according to the data to be updated.
2. The cockpit command recognition device of claim 1, further comprising,
and the keyboard mouse and screen display module is used for selectively switching the instruction recognition device to a model updating state or a non-model updating state.
3. The cockpit command recognition device of claim 2 further comprising,
and the power supply module is used for supplying power to the voice acquisition module, the English instruction identification module, the instruction input and output functional module, the audio data storage module, the model updating module and the keyboard, mouse and screen display module.
4. A method for identifying an airplane cockpit instruction aiming at a Basstein accent English is characterized by comprising the following steps,
a step S1 of providing the instruction recognition device of any one of claims 1 to 3;
step S2, collecting English instruction voice provided by the outside by the voice collecting module and generating English instruction audio signal corresponding to the English instruction voice;
step S3, receiving the English instruction audio signal from the voice acquisition module by an English instruction recognition module, and matching a corresponding English instruction recognition result in a built-in English instruction recognition engine according to the English instruction audio signal; and the number of the first and second groups,
step S4, the input/output module receives the english instruction recognition result from the english instruction recognition module and matches the corresponding instruction code in the built-in instruction code query engine according to the english instruction recognition result to output the external instruction.
5. A method for identifying an airplane cockpit instruction aiming at a Basstein accent English is characterized by comprising the following steps,
a step S1 of providing the instruction recognition device of any one of claims 1 to 3;
step S2, collecting English instruction voice provided by the outside by the voice collecting module and generating English instruction audio signal corresponding to the English instruction voice;
step S3, the English instruction recognition module receives the English instruction audio signal from the voice acquisition module and matches the English instruction recognition result corresponding to the English instruction audio signal in a built-in English instruction recognition engine according to the English instruction audio signal;
step S4, the audio data storage module receives the English instruction audio signal from the voice acquisition module and the English instruction recognition result from the English instruction recognition module, and in a model updating state, the English instruction audio signal and the English instruction recognition result are judged to be stored or discarded according to recognition confidence, so that data to be updated are generated; and the number of the first and second groups,
step S5, the model updating module receives the data to be updated from the audio data storage module in the model updating state and updates the matching model used by the english instruction recognition engine built in the english instruction recognition module according to the data to be updated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810781122.7A CN110797009A (en) | 2018-07-17 | 2018-07-17 | Aircraft cabin instruction recognition device to bakistan accent english |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810781122.7A CN110797009A (en) | 2018-07-17 | 2018-07-17 | Aircraft cabin instruction recognition device to bakistan accent english |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110797009A true CN110797009A (en) | 2020-02-14 |
Family
ID=69424958
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810781122.7A Pending CN110797009A (en) | 2018-07-17 | 2018-07-17 | Aircraft cabin instruction recognition device to bakistan accent english |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110797009A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2642482A1 (en) * | 2012-03-23 | 2013-09-25 | Tata Consultancy Services Limited | Speech processing method and system adapted to non-native speaker pronunciation |
| CN103632668A (en) * | 2012-08-21 | 2014-03-12 | 北京百度网讯科技有限公司 | Method and apparatus for training English voice model based on Chinese voice information |
| CN104575516A (en) * | 2013-10-07 | 2015-04-29 | 霍尼韦尔国际公司 | Systems and methods for correcting accent-induced speech in an aircraft cockpit using a dynamic speech database |
| CN104570835A (en) * | 2014-12-02 | 2015-04-29 | 苏州长风航空电子有限公司 | Cockpit voice command control system and operating method thereof |
| CN105408952A (en) * | 2013-02-21 | 2016-03-16 | 谷歌技术控股有限责任公司 | Recognizing accented speech |
| US20170256254A1 (en) * | 2016-03-04 | 2017-09-07 | Microsoft Technology Licensing, Llc | Modular deep learning model |
| US9786271B1 (en) * | 2016-09-28 | 2017-10-10 | International Business Machines Corporation | Voice pattern coding sequence and cataloging voice matching system |
-
2018
- 2018-07-17 CN CN201810781122.7A patent/CN110797009A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2642482A1 (en) * | 2012-03-23 | 2013-09-25 | Tata Consultancy Services Limited | Speech processing method and system adapted to non-native speaker pronunciation |
| CN103632668A (en) * | 2012-08-21 | 2014-03-12 | 北京百度网讯科技有限公司 | Method and apparatus for training English voice model based on Chinese voice information |
| CN105408952A (en) * | 2013-02-21 | 2016-03-16 | 谷歌技术控股有限责任公司 | Recognizing accented speech |
| CN104575516A (en) * | 2013-10-07 | 2015-04-29 | 霍尼韦尔国际公司 | Systems and methods for correcting accent-induced speech in an aircraft cockpit using a dynamic speech database |
| CN104570835A (en) * | 2014-12-02 | 2015-04-29 | 苏州长风航空电子有限公司 | Cockpit voice command control system and operating method thereof |
| US20170256254A1 (en) * | 2016-03-04 | 2017-09-07 | Microsoft Technology Licensing, Llc | Modular deep learning model |
| US9786271B1 (en) * | 2016-09-28 | 2017-10-10 | International Business Machines Corporation | Voice pattern coding sequence and cataloging voice matching system |
Non-Patent Citations (1)
| Title |
|---|
| 苏意玲等: "英语作为第二语言的多媒体语音数据库设计制作及初步测试", 《贵州大学学报(自然科学版)》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108520743B (en) | Voice control method of intelligent device, intelligent device and computer readable medium | |
| EP0840288B1 (en) | Method and system for editing phrases during continuous speech recognition | |
| US6363347B1 (en) | Method and system for displaying a variable number of alternative words during speech recognition | |
| CN105501121B (en) | A kind of intelligence awakening method and system | |
| US5829000A (en) | Method and system for correcting misrecognized spoken words or phrases | |
| CN107301865B (en) | Method and device for determining interactive text in voice input | |
| US5899976A (en) | Method and system for buffering recognized words during speech recognition | |
| KR102007478B1 (en) | Device and method for controlling application using speech recognition under predetermined condition | |
| CN206595039U (en) | A kind of interactive system for vehicle-mounted voice | |
| EP1346343A1 (en) | Speech recognition using word-in-phrase command | |
| CN109243462A (en) | Voice awakening method and device | |
| WO2001084535A2 (en) | Error correction in speech recognition | |
| US10714082B2 (en) | Information processing apparatus, information processing method, and program | |
| JPWO2016103415A1 (en) | Head mounted display system and method for operating head mounted display device | |
| JP2016521383A (en) | Method, apparatus and computer readable recording medium for improving a set of at least one semantic unit | |
| CN115410572A (en) | Voice interaction method, device, terminal, storage medium and program product | |
| US10847154B2 (en) | Information processing device, information processing method, and program | |
| CN104424942A (en) | Method for improving character speed input accuracy | |
| KR101385012B1 (en) | Multi-modal input device using handwriting and voice recognition and control method thereof | |
| CN110797009A (en) | Aircraft cabin instruction recognition device to bakistan accent english | |
| KR101250897B1 (en) | Apparatus for word entry searching in a portable electronic dictionary and method thereof | |
| EP0840287A2 (en) | Method and system for selecting recognized words when correcting recognized speech | |
| CN111261164A (en) | Unmanned aerial vehicle speech recognition controlling means | |
| US20220100959A1 (en) | Conversation support device, conversation support system, conversation support method, and storage medium | |
| CN112102812B (en) | Anti-false wake-up method based on multiple acoustic models |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200214 |