KR20160066347A - Apparatus and method for recognizing voice in vehicle - Google Patents

Abstract

The present invention suggests an apparatus and a method for recognizing voice in a vehicle to recognize the voice of a user by removing a beep sound from sound information inputted to a micro phone and by using customized database for the user. The device according to the present invention includes a beep sound removing part which removes the beep sound from the sound information and obtains the voice of the user when the sound information is inputted by the user in the vehicle; a user information obtaining part which obtains information about the user; and a vocabulary list detecting part which detects a vocabulary list from the database based on the information about the user; and a voice recognition part which recognizes the voice of the user based on the vocabulary list.

Description

[0001] Apparatus and method for recognizing voice in vehicle [

The present invention relates to a speech recognition apparatus and method. More particularly, the present invention relates to a vehicle speech recognition apparatus and method.

Generally, many devices are installed in the vehicle, and devices for the convenience of the driver and the passengers are constantly being developed and installed. Typical examples are audio devices and air conditioners, and moreover, there is an increasing trend of vehicles equipped with navigation systems that guide the route to the destination by locating the vehicle through a satellite or the like.

However, it is true that such convenience devices provide various convenience to the driver and the passengers. However, there is a problem that the driver has a low concentration of driving in the process of operating the convenience devices while driving, .

In order to solve such a problem, in recent years, a vehicle voice recognition device has been developed in which a driver recognizes a voice of a driver and controls various convenience devices according to a voice command of the driver when the driver commands the voice without directly manipulating the convenience devices.

However, since only a microphone for recognizing the driver's voice exists in the conventional vehicle voice recognition device, the noise generated from the inside or the outside of the vehicle is inputted to the microphone together with the voice, so that the voice control is not properly performed.

Korean Patent Laid-Open Publication No. 2013-0063091 proposes a speech recognition device for a vehicle. However, this apparatus can not solve the above-described problem because it is selectively activated as needed to recognize the voice command of the user.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a vehicle voice recognition apparatus and method for recognizing voice of a user by removing a beep from sound information input by a microphone and using a database specialized for the user .

However, the objects of the present invention are not limited to those mentioned above, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

SUMMARY OF THE INVENTION The present invention is conceived to achieve the above-mentioned object, and it is an object of the present invention to provide a beep tone remover which removes a beep sound from the sound information when a sound information is input by a user in a vehicle, A user information acquisition unit for acquiring information about the user; A vocabulary list detection unit for detecting a vocabulary list from a database based on information about the user; And a voice recognition unit for recognizing the voice of the user based on the lexical list.

Preferably, the beep canceller removes noise from the sound information in addition to the beep sound.

Preferably, the beep canceling unit removes the beep using an echo canceller, and removes the noise using noise reduction.

Preferably, the beep canceler uses an echo canceller including an adaptive filter, a residual echo suppressor, and a comfort noise generator as the echo canceller.

Preferably, the beep canceller removes the beep first from the sound information and removes the noise later.

Preferably, the beep canceller converts a stereo component to a mono component with respect to the audio data included in the sound information, and uses the down sampling to equalize the frequency of each component included in the sound information And removes the beep sound from the sound information.

Preferably, the lexical item list detector collects and collects at least one of words, sentences, and commands written by the user as a speech uttered over the reference number of times from the database, and detects the lexical list.

Preferably, the vehicle voice recognition apparatus is driven when a speech sound is input from the user after the beep sound is reproduced.

In addition, the present invention provides a method of controlling a mobile terminal, comprising the steps of: receiving sound information from a user in a vehicle, removing a beep sound from the sound information; Obtaining information about the user; Detecting a vocabulary list from a database based on information about the user; And recognizing the voice of the user based on the lexical list.

Advantageously, the step of acquiring the user's voice also removes noise from the sound information, in addition to the beep sound.

Preferably, the step of acquiring the user's voice removes the beep using an echo canceller, and removes the noise using noise reduction.

Preferably, the step of acquiring the voice of the user further comprises an echo canceller including an adaptive filter, a residual echo suppressor, and a comfort noise generator as the echo canceller .

Advantageously, the step of acquiring the user's voice first removes the beep from the sound information and removes the noise later.

Preferably, the step of acquiring the voice of the user may include converting the stereo component to mono component with respect to the audio data included in the sound information, and converting the stereo component into mono component using down sampling, The same frequency is converted, and the beep sound is removed from the sound information.

Preferably, the detecting includes collecting and collecting at least one of a word, a sentence, and a command written by the user as an utterance exceeding a reference number of times from the database, and detects the word list as the vocabulary list.

Preferably, the vehicle speech recognition method is performed when a speech sound is input from the user after the beep sound is reproduced.

The present invention can obtain the following effects by removing the beep sound from the sound information input by the microphone and recognizing the user's voice by using a database specialized for the user.

First, in case of an utterance of an instruction, even if it is mixed with a beep sound, the recognition operation can be performed, thereby improving the user's perceived performance. In addition, by optimizing recognition commands for each user, optimized recognition performance can be improved for users.

Second, speech recognition users are more likely to recognize speech even if they are misaligned in the case of a user who is a beginner of speech recognition or a user who is in a hurry. In addition, user-optimized usability can be presented by improving the recognition performance mainly on the commands used mainly by the user.

Third, while the timing of the existing ignition is limited, the timing of the ignition can be changed, so that the timing system suitable for the user can be designed. Also, it is possible to deduce what functions frequently used by the user when the list of commands for each user is provided.

Fourthly, not only the beep but also the surrounding noise canceling is carried out to transmit a clean voice rather than a complex voice to the voice recognition engine.

Fifth, recognition speed can be improved by providing a list corresponding to a user when recognizing an instruction.

Sixth, when a clean voice is transmitted to the engine, the reliability of recognition operation is improved.

Seventh, it is economical and not harmful to the environment because it is developed on a system using eco-canler as a SW.

1 is a conceptual diagram illustrating a speech preprocessing system according to an embodiment of the present invention.
FIG. 2 is a conceptual diagram showing a SW layer diagram of the speech preprocessing system shown in FIG. 1. FIG.
FIG. 3 is a diagram comparing the analysis result of a signal inputted with a microphone and a signal obtained by removing echo.
4 is an operational block diagram of the echo canceller constituting the speech preprocessing system.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to designate the same or similar components throughout the drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. In addition, the preferred embodiments of the present invention will be described below, but it is needless to say that the technical idea of the present invention is not limited thereto and can be variously modified by those skilled in the art.

Speech recognition is an item that requires an emotional approach rather than a logical one. It is often the case that a user feels emotional complaints when he or she feels it is hard to recognize after use or difficult to use. In other words, the customer feels that there are many problems with the feeling index.

In the case of complaining about usability of speech recognition among such IQS user complaints, a large number of users complain about the mixed voice of various noise and beep sound as well as the user's voice in the voice information input by the microphone.

In addition, in the conventional speech recognition system, the commands used by the user in the entire speech recognition commands are limited, but the recognition rate is limited due to the competition of recognition of the entire commands in the same condition every time.

Disclosure of Invention Technical Problem [8] Accordingly, the present invention has been made to solve the above problems, and it is an object of the present invention to provide a system for eliminating a beep sound from a user's voice by using an echo canceller and recognizing a user's voice using a database specialized for the user.

Hereinafter, a more detailed description will be given with reference to the drawings.

1 is a conceptual diagram illustrating a speech preprocessing system according to an embodiment of the present invention.

The speech preprocessing system 100 according to the present invention is a preprocessing system for removing a beep sound from a speech of a speaking person using an echo canceller 125. In other words, the speech preprocessing system 100 uses the echo canceller 125 to remove the beep sound from the speaker's voice at the time of the speech together with the speech recognition beep by the novice speaker, and then transmits the beep sound to the recognition engine 150. The speech preprocessing system 100 improves the speech recognition usability by constructing a recognition DB based on instructions that are well used for each user by classifying the speakers using the speech characteristics of each speaker.

1, the speech preprocessing system 100 includes a microphone 105, an audio data buffer 110, a type conversion unit 115, a sampling unit 120, an echo canceller 125, a noise reduction unit (not shown) A speaker identification instruction word list construction unit 140, a speaker recognition instruction word list calling unit 145 and a recognition engine 150. The speaker recognition instruction list construction unit 140,

The microphone 105 performs a function of receiving voice from a driver. At this time, beep sound, ambient noise, etc., along with the driver's voice, are input to the microphone 105 and transmitted to the audio data buffer 110.

The audio data buffer 110 temporarily stores the sound information mixed with the driver's voice, beep sound, ambient noise, etc. in the form of digital data. Then, the audio data buffer 110 transmits the sound information to the type conversion unit 115.

The type conversion unit 115 performs a function of converting a stereo characteristic into a mono characteristic with respect to audio data (e.g., driver's voice) constituting the sound information.

The sampling unit 120 performs a function of converting the frequency of each component constituting the sound information equally. The sampling unit 120 converts the frequencies of the driver's voice and beep sound to 16 KHz by using a down sampling technique.

The echo canceller 125 performs a function of removing the beep sound from the sound information that has passed through the type conversion unit 115 and the sampling unit 120. [

The noise reduction 130 performs a function of removing noise from the sound information from which the beep sound is removed. In this embodiment, the echo canceller 125 first removes the beep sound from the sound information, and then the noise reduction 130 eliminates the noise, but is not necessarily limited to this, and the noise reduction 130 may be an echo canceller It is also possible to perform the function before the step 125.

The recognition engine 150 performs a function of receiving the clean speech remaining after the beep and noise are removed from the sound information, that is, the pure voice of the driver, from the noise reduction 130. When the driver's voice is received by the recognition engine 150, the speaker determination unit 135, the speaker recognition instruction word list construction unit 140, the speaker recognition instruction list calling unit 145, and the like are sequentially driven.

The recognition engine 150 generates a list of recognizing commands based on the list of per-speaker recognition commands obtained from the sequential driving of the speaker determining unit 135, the per-speaker recognition command list construction unit 140, and the per-speaker recognition command list calling unit 145, And performs a voice recognition function.

The speaker discrimination unit 135 receives personal information such as an ID from the driver when the driver's pure voice is input to the recognition engine 150, and performs a function of discriminating a speaker.

The per-speaker recognition command list constructing unit 140 detects a command from the database when the speaker is discriminated by the speaker discriminating unit 135, and constructs a list of the detected commands.

The per-speaker recognition instruction list calling unit 145 performs a function of transmitting the list generated by the per-speaker recognition instruction list construction unit 140 to the recognition engine 150.

FIG. 2 is a conceptual diagram showing a SW layer diagram of the speech preprocessing system shown in FIG. 1. FIG. In FIG. 2, a method for optimizing speech recognition for each speaker based on a pure voice of a speaker (driver) whose beep sound, noise, etc. have been removed and a list of speech recognition instruction words by speaker will be described. The following description refers to Fig. 1 and Fig.

The audio data buffer 110 receives sound information such as a speech recognition beep sound and surrounding noise in addition to the speaker voice inputted to the microphone (MIC) 105 through an audio data buffer management layer (an audio data buffer management) The audio data buffer 110 receives sound information in the form of audio data through an audio driver (Codec) (Audio Driver) 240.

Then, the type conversion unit 115 converts the sound of the speaker, the beep sound, etc. constituting the sound information from the stereo type to the mono type through the first layer 220 and the second layer 230.

Then, the sampling unit 120 down-samples the sampling frequency of the beep sound from 44.1 KHz to 16 KHz through the third layer 225.

Thereafter, the echo canceller 125 removes the beep sound from the speech sound according to the existing beep frequency characteristics in the audio data including the surrounding noise through the echo canceling layer 215.

On the other hand, the noise reduction 130 removes surrounding noise existing after the beep sound is removed through the echo cancellation layer 215.

Then, the recognition engine 150 receives the beep and the noise-removed speech data through the OS framework layer 210.

Then, the speaker determination unit 135 determines whether the speaker is an existing speaker or a new speaker based on the speaker characteristics from the recognition engine 150 through the application layer 205.

Then, when the per-speaker recognition command list construction unit 140 is a new speaker through the application layer 205, a new speaker's voice recognition list is newly created, and in the case of an existing speaker, a voice recognition list corresponding to the speaker is retrieved.

When the recognition engine 150 receives the voice recognition list from the per-speaker recognition command list calling unit 145, the recognition engine 150 recognizes the voice recognition list in order of priority through the application layer 205.

FIG. 3 is a diagram comparing the analysis result of a signal inputted with a microphone and a signal obtained by removing echo.

3 (a) shows the analysis result of the signal input to the microphone. In FIG. 3A, reference numeral 310 denotes a TTS prompt prompt interval (about 1.8 seconds), and reference numeral 320 denotes a section 320 in which a dial number 321 and a beep sound 322 are mixed .

As shown in FIG. 3 (a), when a signal input to the microphone is analyzed, a beep sound 322 exists in the middle of the speaker's voice 321, resulting in a failure in speech recognition.

FIG. 3 (b) shows the result of analysis of the echo canceled signal. In FIG. 3 (b), reference numeral 330 denotes a filter convergence period (about 0.5 second), and reference numeral 350 denotes a recognizable period to be transmitted to the recognition engine. Reference numeral 340 denotes an extendable section of the recognizable section.

As shown in FIG. 3 (b), when the present invention is applied, it is confirmed that only the pure speech of the speaker is detected by removing the beep sound by echo cancellation.

4 is an operational block diagram of the echo canceller constituting the speech preprocessing system.

The MIC 405 receives a user's utterance (s (n): utterance). At this time, an echo signal d (n): echo including a beep sound is also input from the SPEAKER 410 to the MIC 405.

The adaptive filter h (n) 425 performs a function of predicting and removing the echo signal through adaptation, in which the sound exiting through the SPEAKER 410 is introduced into the MIC 405 again in the form of an echo .

The detection and control unit 420 performs a double talk detection function in a signal input to the MIC 405. [ The detection and control unit 420 also performs an adaptation filter 425 for adjusting the adaptation rate so that the adaptation filter 425 can predict and remove the echo signal through adaptation.

The Residual Echo Suppressor 415 performs the function of eliminating the echo that has not been sufficiently removed by the adaptive filter 425.

Due to the echo cancellation, there is a silence interval, which causes the noise characteristics to be inconsistent. The comfort noise generator 430 performs a function of generating and adding comfort noise similar to background noise to prevent this.

If a novice user or an intermediate user who is using voice recognition for the first time is unsteady or unfamiliar, when the voice recognition command is uttered, the user should normally listen to the beep sound and recognize it normally.

However, speech recognition fails when speech is heard only by the announcement without waiting for the beep sound. In other words, the voice recognition user does not blame the immaturity of use and complains that the system does not recognize it.

The following are summarized as follows.

Use Case 1: When the beep sound is reproduced → Recognition succeeded

Use Case 2: In case of ignition before beep sound reproduction → However, it is difficult to occur because interval time between microphone open time and beep sound reproduction is as short as 30 ms → recognition failure

Use Case 3: In case of ignition during beep sound → recognition failure

In the case of the present invention, the use of the present invention can be realized even in Use Cases 2 and 3 by applying the present invention to the time between the announcement moment playback and the beep playback.

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention has been described with reference to Figs. Best Mode for Carrying Out the Invention Hereinafter, preferred forms of the present invention that can be inferred from the above embodiment will be described.

The speech recognition apparatus for a vehicle according to a preferred embodiment of the present invention includes a beep sound removing unit, a user information obtaining unit, a lexical word list detecting unit, a speech recognizing unit, a power source unit, and a main control unit.

The power supply unit performs a function of supplying power to each configuration of the voice recognition device for a vehicle. The main control unit performs a function of controlling the overall operation of each component constituting the vehicle voice recognition apparatus. Considering that the vehicle voice recognition device is provided in the vehicle, the power supply unit and the main control unit can be replaced by the vehicle battery and the MCU (or ECU), so that the power supply unit and the main control unit may not be provided in the present embodiment.

The beep canceller performs a function of removing a beep sound from sound information and acquiring a user's voice when the sound information is input by the user in the vehicle.

The beep canceller can remove noise from the sound information in addition to the beep sound.

The beep canceller removes the beep using an echo canceller, and noise can be removed using noise reduction.

The beep canceller may use an echo canceller including an adaptive filter, a residual echo suppressor, and a comfort noise generator as the echo canceller.

The beep canceller may first remove the beep from the sound information and remove the noise later.

The beep canceller converts a stereo component into a mono component with respect to the audio data included in the sound information, converts the frequency of each component included in the sound information equally by using down sampling, The sound can be removed.

The user information obtaining unit performs a function of obtaining information about the user.

The vocabulary list detection unit performs a function of detecting a vocabulary list from the database based on information about the user.

The vocabulary list detection unit may collect and collect at least one of words, sentences, and commands recorded by the user as a speech that has been uttered more than the reference number of times, and may detect the vocabulary list. For this purpose, the vocabulary list detecting unit may increase the number of utterances for the corresponding vocabulary by one every time the utterance voice of the user is recognized.

The speech recognition unit performs a function of recognizing the user's speech based on the vocabulary list.

The vehicle voice recognition apparatus described above can be driven when a voice is input from the user after the beep sound is reproduced.

Next, an operation method of the vehicle voice recognition apparatus will be described.

First, when the beep sound removing unit inputs sound information by a user on the vehicle, the beep sound is removed from the sound information to acquire the user's voice.

Then, the user information acquisition unit acquires information about the user.

Then, the lexical-list detecting unit detects the lexical-list from the database based on the information about the user.

Then, the speech recognition unit recognizes the user's speech based on the lexical list.

It is to be understood that the present invention is not limited to these embodiments, and all elements constituting the embodiment of the present invention described above are described as being combined or operated in one operation. That is, within the scope of the present invention, all of the components may be selectively coupled to one or more of them. In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. In addition, such a computer program may be stored in a computer readable medium such as a USB memory, a CD disk, a flash memory, etc., and read and executed by a computer to implement an embodiment of the present invention. As the recording medium of the computer program, a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like can be included.

Furthermore, all terms including technical or scientific terms have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined in the Detailed Description. Commonly used terms, such as predefined terms, should be interpreted to be consistent with the contextual meanings of the related art, and are not to be construed as ideal or overly formal, unless expressly defined to the contrary.

It will be apparent to those skilled in the art that various modifications, substitutions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims. will be. Therefore, the embodiments disclosed in the present invention and the accompanying drawings are intended to illustrate and not to limit the technical spirit of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments and the accompanying drawings . The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

Claims (13)

A beep sound removing unit for removing a beep sound from the sound information when the sound information is input by the user in the vehicle to acquire the voice of the user;
A user information acquisition unit for acquiring information about the user;
A vocabulary list detection unit for detecting a vocabulary list from a database based on information about the user; And
A voice recognition unit for recognizing the voice of the user based on the vocabulary list,
And a voice recognition device for recognizing the voice of the vehicle. The method according to claim 1,
And the beep canceller removes noise from the sound information in addition to the beep sound. 3. The method of claim 2,
Wherein the beep canceller removes the beep by using an echo canceller and removes the noise by using noise reduction. The method of claim 3,
Wherein the beep canceller uses an echo canceller including an adaptive filter, a residual echo suppressor, and a comfort noise generator as the echo canceller, Device. 3. The method of claim 2,
Wherein the beep canceller removes the beep first from the sound information and removes the noise later. The method according to claim 1,
The beep canceller may convert a stereo component to a mono component with respect to the audio data included in the sound information, convert the frequency of each component included in the sound information by using down sampling, And the beep sound is removed from the sound information. The method according to claim 1,
Wherein the lexical-list detecting unit collects and collects at least one of words, sentences, and commands recorded by the user as being uttered over a reference number of times from the database, and detects the lexical list. The method according to claim 1,
Wherein the vehicle speech recognition apparatus is driven when a speech sound is input from the user after the beep sound is reproduced. Removing a beep sound from the sound information when the sound information is input by the user in the vehicle, thereby obtaining the user's voice;
Obtaining information about the user;
Detecting a vocabulary list from a database based on information about the user; And
Recognizing the voice of the user based on the vocabulary list
And a voice recognition unit for recognizing the voice of the vehicle. 10. The method of claim 9,
Wherein the step of acquiring the voice of the user removes noise from the sound information in addition to the beep sound. 11. The method of claim 10,
Wherein the step of acquiring the voice of the user removes the beep using an echo canceller and removes the noise using noise reduction. 12. The method of claim 11,
Wherein the step of acquiring the voice of the user uses an echo canceller including an adaptive filter, a residual echo suppressor, and a comfort noise generator as the echo canceller, A voice recognition method for a vehicle. 11. The method of claim 10,
Wherein the step of acquiring the voice of the user removes the beep first from the sound information and removes the noise later.

Images