KR102474804B1 - Apparatus for controlling voice recognition, system having the same and method thereof - Google Patents

Abstract

The present invention relates to a voice recognition control device, a system including the same, and a method thereof. A voice recognition control device according to an embodiment of the present invention includes a voice signal input unit for receiving a voice signal including a wakeup keyword and a command; a voice filtering unit that recognizes a wakeup keyword located at the head of the voice signal and wakes up a device corresponding to the wakeup keyword among a plurality of devices; and a command filtering unit that selects main commands based on words and verbs of commands of the voice signal.

Description

Apparatus for controlling voice recognition, system having the same and method thereof}

The present invention relates to a voice recognition control device, a system including the same, and a method thereof, and more particularly, to a technology for performing accurate voice command control in an effective device among a plurality of devices.

As devices become smarter, a voice recognition function capable of executing a device function using a user's voice signal is being installed in the device.

In order to use the voice recognition function built into the device, the voice recognition function of the device must be woken up. Accordingly, a wake-up is performed using a fixed wake-up keyword. Due to this, when a plurality of devices equipped with the same voice recognition function exist in the same place, the voice recognition function of an unwanted device can be woken up.

In addition, the existing voice recognition function separately processes a wake-up keyword and a voice command. Accordingly, after the user inputs the wakeup keyword, when the voice recognition function of the device wakes up, the user must input a voice command. If a user continuously inputs a wake-up keyword and a voice command, the existing voice recognition function may not wake up, or even if the user wakes up, an error in recognizing the input voice command may occur.

Accordingly, there is a demand for a technology capable of more accurately recognizing a voice command while more conveniently and accurately waking up a voice recognition function of a device.

An embodiment of the present invention is to provide a voice recognition control device, a system including the same, and a method for controlling a vehicle through accurate voice recognition in an effective device among a plurality of devices.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description below.

A voice recognition control device according to an embodiment of the present invention includes a voice signal input unit for receiving a voice signal including a wakeup keyword and a command; a voice filtering unit that recognizes a wakeup keyword located at the head of the voice signal and wakes up a device corresponding to the wakeup keyword among a plurality of devices; and a command filtering unit that selects main commands based on words and verbs of commands of the voice signal.

In one embodiment, a verification unit for verifying whether to transmit a vehicle command through the wake-up device may be further included.

In one embodiment, the verifier checks whether the command of the audio signal conforms to the command grammar for each device when the number of the wake-up devices is at least one, and the command of the audio signal is the at least one wake-up device. It may include selecting a device having the highest grammar matching rate with and transmitting the vehicle command.

In one embodiment, the voice filtering unit extracts text from the voice signal, and when at least one wakeup keyword exists in the voice signal, voice signal correlation and text correlation with respect to the at least one wakeup keyword It may include selecting a wakeup keyword having a high correlation by performing .

In one embodiment, when a plurality of voice signals are input, the command filtering unit determines a character matching rate and signal strength of each command, and selects a voice signal having a high character matching rate or a large signal strength. may include doing

In one embodiment, the command filtering unit may include selecting one of the plurality of commands based on the number of verbs and words of the command when the voice signal includes a plurality of commands.

In one embodiment, the command filtering unit may include selecting a command having a large number of verbs and words in the command, and selecting the command by prioritizing the number of verbs over the number of words.

In one embodiment, the biometric authentication unit may further include a biometric authentication unit performing biometric authentication when the voice signal is input.

In one embodiment, when verification is completed by the verification unit, a feedback unit for feeding back that the wakeup keyword has been recognized to the user who uttered the voice signal may be further included.

In one embodiment, the voice filtering unit may include determining that the wakeup keyword is a valid wakeup keyword when the wakeup keyword is located at the beginning of the voice signal when the wakeup keyword exists in the voice signal.

A voice recognition control system according to an embodiment of the present invention recognizes a wakeup keyword located at the head of a user's voice signal for a plurality of devices having different wakeup keywords, and selects a device corresponding to the recognized wakeup keyword. a voice recognition control device that wakes up and filters commands in the voice signal; and a voice recognition server that performs voice recognition on the voice signal received from the device and provides a result of the voice recognition to the device.

In one embodiment, the device may include a voice signal input unit for receiving a voice signal including a wakeup keyword and a command; a voice filtering unit that recognizes a wakeup keyword located at the head of the voice signal and wakes up a device corresponding to the wakeup keyword among a plurality of devices; a command filtering unit for selecting main commands based on words and verbs of commands of the voice signal; and a verification unit verifying whether to transmit a vehicle command through the wake-up device.

In one embodiment, the verifier checks whether the command of the audio signal conforms to the command grammar for each device when the number of the wake-up devices is at least one, and the command of the audio signal is the at least one wake-up device. It may include selecting a device having the highest grammar matching rate with and transmitting the vehicle command.

In one embodiment, the voice filtering unit extracts text from the voice signal, and when at least one wakeup keyword exists in the voice signal, voice signal correlation and text correlation with respect to the at least one wakeup keyword It may include selecting a wakeup keyword having a high correlation by performing .

In one embodiment, when a plurality of voice signals are input, the command filtering unit determines a character matching rate and signal strength of each command, and selects a voice signal having a high character matching rate or a large signal strength. may include doing

In one embodiment, the command filtering unit may include selecting one of the plurality of commands based on the number of verbs and words of the command when the voice signal includes a plurality of commands.

A voice recognition control method according to an embodiment of the present invention includes receiving a voice signal including a wakeup keyword and a command; recognizing a wakeup keyword located at the head of the voice signal and waking up a device corresponding to the wakeup keyword among a plurality of devices; selecting a main command based on words and verbs of the command of the voice signal; verifying whether to transmit a vehicle command through the wake-up device; and transmitting a vehicle command through the verified device.

In one embodiment, the verifying step checks whether the command of the voice signal conforms to the command grammar for each device, and selects a device with the highest grammar matching rate of the command of the voice signal with the at least one wake-up device. It may include transmitting a vehicle command by doing so.

In one embodiment, the step of waking up the device may include extracting text from the voice signal, and if at least one wakeup keyword exists in the voice signal, voice signal correlation with the at least one wakeup keyword and selecting a wakeup keyword having a high correlation by performing text correlation.

In one embodiment, in the step of selecting the main command, when a plurality of voice signals are input, a character matching rate and signal strength of each command are determined, and the character matching rate is high or the signal strength is high. It may include selecting an audio signal.

It is possible to perform vehicle control through accurate voice recognition in an effective device among a plurality of devices according to an embodiment of the present invention.

1 is a block diagram of a voice recognition control system according to an embodiment of the present invention.
2 is a detailed configuration diagram of the device of FIG. 1;
3 is a diagram for explaining a voice filtering method according to an embodiment of the present invention.
4 is a diagram showing indexes according to correlation of voice command signals according to an embodiment of the present invention.
5 is a detailed configuration diagram of the voice recognition server of FIG. 1 .
6 is a flowchart illustrating a multi-dialogue platform voice recognition control method according to an embodiment of the present invention.
7 is a flowchart illustrating a voice filtering method in detail according to an embodiment of the present invention.
8 is a flowchart specifically illustrating a command filtering method according to an embodiment of the present invention.
9 is a flowchart illustrating a verification method in detail according to an embodiment of the present invention.
10 is a diagram for explaining a voice recognition control method according to an embodiment of the present invention.
11 is a configuration diagram of a computer system to which a voice recognition control method according to an embodiment of the present invention is applied.

Hereinafter, some embodiments of the present invention will be described in detail through exemplary drawings. In adding reference numerals to components of each drawing, it should be noted that the same components have the same numerals as much as possible even if they are displayed on different drawings. In addition, in describing an embodiment of the present invention, if it is determined that a detailed description of a related known configuration or function hinders understanding of the embodiment of the present invention, the detailed description will be omitted.

In describing the components of the embodiment of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are only used to distinguish the component from other components, and the nature, order, or order of the corresponding component is not limited by the term. In addition, unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art to which the present invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in the present application, they should not be interpreted in an ideal or excessively formal meaning. don't

Hereinafter, embodiments of the present invention will be described in detail with reference to FIGS. 1 to 11 .

1 is a block diagram of a voice recognition control system according to an embodiment of the present invention.

Referring to FIG. 1, the voice recognition control system according to an embodiment of the present invention includes a voice recognition control device 100 to which a voice signal is input, at least one device 301, .... 300n, and a voice recognition control device 100. The device 300 transmits a command recognizing the voice signal received from the device 300, and the device 300 transmits the command to the vehicle terminal 400.

The voice recognition control apparatus 100 recognizes a wakeup keyword located at the head of a user's voice signal for a plurality of devices having different wakeup keywords and wakes up a device corresponding to the recognized wakeup keyword, Commands in the voice signal are filtered, the wake-up device is verified, and the vehicle command is transmitted to the vehicle terminal 400 through the verified wake-up device.

The voice recognition server 200 performs voice recognition upon receiving a voice signal from the voice recognition control device 100 . That is, the voice recognition server 200 converts a voice signal into a character string (text), and may include an acoustic model and a language model for this. The acoustic model refers to modeling the signal characteristics of voice. The language model refers to modeling a linguistic sequence relationship of words or syllables corresponding to a recognized vocabulary.

At least one or more devices 301, ..., 30n recognize different wakeup keywords, and may include a vehicle terminal in a vehicle, a mobile communication terminal, and the like. The vehicle terminal may include all devices mounted in the vehicle, such as audio, video navigation (AVN) having a voice recognition function, a telematics terminal, and a navigation device. Mobile communication terminals include smart phones, notebooks, smart boards, tablet PCs (personal computers), handheld devices, handheld computers, media players, and e-book devices. , And at least one of voice recognition devices such as PDA (Personal Digital Assistant), but in the present disclosure, the device 300 is not limited to the above.

At this time, the voice recognition control device 100, at least one or more devices 301, .... 300n, and the voice recognition server 200 may be connected to a wired or / and wireless network-based, a short-range wireless network, or / and a long-distance wireless network. have.

FIG. 2 is a detailed configuration diagram of the voice recognition control device 100 of FIG. 1 .

The voice recognition control device 100 receives a voice signal from the user, wakes up when a wakeup keyword is recognized, filters and verifies commands of the voice signal, and transmits the voice signal to the voice recognition server 200 .

To this end, the voice recognition control device 100 includes a voice signal input unit 111, a communication unit 112, a display unit 113, a storage unit 114, a voice filtering unit 115, a command filtering unit 116, and a verification unit. It includes a unit 117, a control unit 118, a biometric authentication unit 119, and a feedback unit 120.

The voice signal input unit 111 receives a voice signal through a microphone or a single input voice device. At this time, the voice signal may include a wakeup keyword and a command.

The communication unit 112 performs communication with the device 300 and the voice recognition server 200 .

The display unit 113 displays the voice recognition result received from the voice recognition server 200 . To this end, the display unit 113 includes a liquid crystal display, a thin film transistor-liquid crystal display, an organic light-emitting diode, a flexible display, and a three-dimensional A display (3D display) or an electrophoretic display (EPD) may be included.

The storage unit 114 stores a voice signal input from a user, a voice recognition result received from the voice recognition server 200, a wakeup keyword list for each device, command grammar information, and the like. To this end, the storage unit 140 may be a flash memory type, a hard disk type, a multimedia card micro type, or a card type memory (for example, SD or XD memory). etc.), RAM (RAM, Random Access Memory), SRAM (Static Random Access Memory), ROM (ROM, Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), It may include a storage medium in the form of a magnetic memory, a magnetic disk, or an optical disk.

The voice filtering unit 115 detects the position of the wakeup keyword among the voice signals input from the user, recognizes the wakeup keyword located at the front end, and wakes up the device when it matches its own wakeup keyword.

To this end, the voice filtering unit 115 extracts a character from the voice signal, filters whether there is a wakeup keyword that exists in the wakeup keyword list, and when multiple wakeup keywords are recognized, each wakeup keyword By performing voice signal correlation and text correlation for the selected wakeup keywords having highly relevant indices, noise is filtered.

3 is a diagram for explaining a voice filtering method according to an embodiment of the present invention. 4 is a diagram showing indexes according to correlation of voice command signals according to an embodiment of the present invention.

Referring to FIG. 3 , the voice signal “Ok, Google. Tell my vehicle to be ready for Alexa? party. When "Gigagenie" is input, voice signal correlation and text correlation are performed for "Ok, Google", "Alexa", and "Gigagenie", respectively. Referring to FIG. 4 , since “Ok, Google” has the highest voice correlation and text correlation value, ok and google having the highest correlation can be selected as a wakeup keyword. At this time, in FIG. 4, the lower the index, the higher the correlation.

At this time, the voice filtering unit 115 selects a command input at the beginning of the voice signal through such correlation and does not select a wakeup keyword input at the middle or end of the voice signal.

The command filtering unit 116 performs text matching on each of a plurality of input voice signals, selects a voice signal having a high matching rate or a high signal strength (SNR), and determines whether the selected voice signal is a single command or multiple commands. judge

In addition, in the case of multiple commands, the command filtering unit 116 identifies the details of the multiple commands, selects the clearest command based on the number of verbs and words, and filters out unclear commands. In this case, the number of verbs may have priority over the number of words.

The verification unit 117 checks the number of devices related to the command, analyzes the command pattern and control path, and determines whether the control path of the command is redundant. If so, check the syntax of the sentence format of the command for the specific device. That is, if there are at least one device that has been woken up, the verification unit 117 checks whether the command of the voice signal conforms to the command grammar for each device, and the command of the audio signal has the highest grammar matching rate with the one or more devices that have been woken up. A high device can be selected to transmit vehicle commands.

The control unit 118 controls the overall operation of each component.

The biometric authentication unit 119 may perform biometric authentication when a voice signal is input.

When verification by the verification unit 117 is completed, the feedback unit 120 feeds back that the wakeup keyword has been recognized to the user who uttered the voice signal.

5 is a detailed configuration diagram of the voice recognition server of FIG. 1 .

The voice recognition server 200 includes a communication unit 210, a storage unit 220, a voice recognition unit 230, and a control unit 240.

The communication unit 210 communicates with the voice recognition control device 100 .

The storage unit 220 stores data for voice recognition.

The voice recognition unit 230 performs voice recognition of the voice signal received from the voice recognition control device 100 .

The controller 240 controls the overall operation of each component.

Although the present invention shows a case in which the voice recognition control device including voice signal filtering, command filtering unit, and verification unit is included in the voice recognition control device 100, it is configured as a terminal separate from a mobile communication terminal or a vehicle terminal. A voice recognition control device may be implemented.

Hereinafter, a voice recognition control method according to an embodiment of the present invention will be described with reference to FIG. 6 . 6 is a flowchart illustrating a voice recognition control method according to an embodiment of the present invention.

Referring to FIG. 6 , the voice recognition control device 100 processes voice filtering on an input voice signal (S100). That is, the voice recognition control apparatus 100 receives a wakeup keyword and a voice signal including a command, recognizes the wakeup keyword located at the head of the voice signal, and wakes up a device corresponding to the wakeup keyword among a plurality of devices. let it

Next, the voice recognition control device 100 performs command filtering (S200). That is, the voice recognition control device 100 selects a main command based on words and verbs of the command of the voice signal.

After that, the voice recognition control device 100 performs a verification process (S300). That is, the voice recognition control apparatus 100 verifies whether to transmit a vehicle command through the wake-up device.

Subsequently, the voice recognition control device 100 transmits a command to the vehicle through the verified device (S400).

Hereinafter, a voice filtering method according to an embodiment of the present invention will be described in detail with reference to FIG. 7 . 7 is a flowchart specifically illustrating a voice filtering method according to an embodiment of the present invention.

The voice recognition control apparatus 100 extracts a character from the input voice signal (S101) and extracts a wakeup keyword from the extracted character (S102). At this time, the wakeup keyword may be located at the beginning, middle, or end of the voice signal, and only the wakeup keyword located at the beginning is recognized as the wakeup keyword in the present invention. For example, if “Hey Siri” is uttered to a device where “Hey Siri” is a wakeup keyword, “Hey Siri” is included in the utterance, but it is not recognized as a wakeup keyword because it is located in the middle of the sentence.

Meanwhile, the voice recognition control device 100 determines whether multiple wakeup keywords are recognized (S103), and if multiple wakeup keywords are not recognized and only a single wakeup keyword is recognized, noise filtering is performed on the corresponding wakeup keyword. After that, it is output (S105).

Meanwhile, when multiple wakeup keywords are recognized, voice signal correlation and text correlation are performed on each wakeup keyword (S105) to select a wakeup command having a highly correlated index (S106).

Subsequently, noise filtering is performed on the selected wakeup command (S105).

Hereinafter, a command filtering method according to an embodiment of the present invention will be described in detail with reference to FIG. 8 . 8 is a flowchart specifically illustrating a command filtering method according to an embodiment of the present invention. 8 describes a filtering method for commands other than the wakeup keyword in the voice signal.

The voice recognition control apparatus 100 waits for input of multiple commands from the device (S201) and determines whether the text matching rate of the multiple commands is 90% or higher (S202). At this time, the text matching rate may be calculated by comparing the text stored in the command DB.

If the text matching rate is not 90% or more, the voice recognition control device 100 selects a command with a high signal strength, and if the text matching rate is 90% or more, selects a command with a high matching rate (S203).

Thereafter, the voice recognition control device 100 determines whether the command is a single command or multiple commands (S205), and transmits the single command to the verifier 117 if it is a single command (S206). In this case, a single command is a case of including one command, and a multi-command is a case of including a plurality of commands. For example, “Open a window and play the radio” is a multi-command.

On the other hand, in the case of multiple commands, the voice recognition control device 100 determines whether the corresponding command is a direct command (S207), and if the command is a direct command, after filtering out unclear commands (S208), the verifier 117 performs a single command is transmitted (S206).

If it is not a direct command, a command is selected based on the number of verbs and words of the command (S209). For example, “start the car” is a direct command, and “prepare to leave soon” is an indirect command. “Start the car” has 1 verb and 3 words, and “Get ready to leave soon” has 1 verb and 4 words. At this time, a command having a large number of verbs or words may be selected.

Thereafter, after performing filtering to remove ambiguous commands for the selected command (S208), a single command is delivered to the verifier 117 (S206).

Hereinafter, a verification method according to an embodiment of the present invention will be described in detail with reference to FIG. 9 . 9 is a flowchart illustrating a verification method in detail according to an embodiment of the present invention.

The voice recognition control apparatus 100 checks the number of wake-up devices for voice recognition (S301), analyzes command patterns and control paths (S302), and determines whether duplicate control paths exist (S303). At this time, the control path refers to a path for transmitting a vehicle command from a wake-up device to a vehicle terminal, and it is determined whether two or more such control paths exist.

If there is no redundant control path, the voice recognition control apparatus 100 selects the corresponding path as an optimal control path (S304), and if there is a redundant control path, checks the grammar for each device of each redundant control path (S305). Accordingly, the grammar matching rate may select a device and transmit a vehicle command to the vehicle terminal.

For example, if device 1 and device 2 wake up, device 1 has a grammar match success rate of 90% and a grammar match failure rate of 10%, and device 2 has a grammar match success rate of 89% and a grammar match failure rate of 5% In case of , device 2 can be selected. At this time, there is a command grammar form for each device, and it is determined whether the grammar matches the grammar of the input command.

Hereinafter, a voice recognition control method according to an example will be described with reference to FIG. 10 . Referring to FIG. 10, when the user says “ Ok, Google . When saying “Tell my vehicle to be ready for the destination of Alexa ’s party.”, the voice filtering unit 115 extracts “Ok, google” and “Alexa” from the wake-up keyword list. “Ok, google” at the beginning of is recognized as a wake-up keyword, and “Alexa” is not recognized as a wake-up keyword. Accordingly, assuming that device 1 and device 2 exist as devices using “Ok, google” as a wake-up keyword, device 1 and device 2 wake up.

After that, the command filtering unit 116 executes the command “Tell my vehicle to be ready for the destination of Alexa’s party.” to filter In other words, since it is a multi-command, if you separate it, “Tell my vehicle to be ready” and “for the destination of Alexa's party” are separated into two commands, and “Tell my vehicle to be ready” is divided into one verb and five words. It is composed of 0 verbs and 6 words “for the destination of Alexa's party”.

Accordingly, the command filtering unit 116 selects “Tell my vehicle to be ready” with a large number of verbs as a main command among multiple commands.

After that, the verification unit 117 identifies control paths transmitted as commands from the two wake-up devices, device 1 and device 2, to the vehicle terminal, and determines the grammar matching rate of the devices. That is, there is a command grammar for each device, and a device with a high grammar matching rate with the “Tell my vehicle to be ready” command uttered by the user is selected. For example, if device 1 has a grammar match success rate of 90% and a grammar match failure rate of 10%, and device 2 has a grammar match success rate of 89% and a grammar match failure rate of 5%, then device 2 with a lower failure rate than the success rate By selecting , a command may be transmitted to the vehicle terminal.

11 is a configuration diagram of a computer system to which a voice recognition control method according to an embodiment of the present invention is applied.

Referring to FIG. 11 , a computing system 1000 includes at least one processor 1100, a memory 1300, a user interface input device 1400, a user interface output device 1500, and a storage connected through a bus 1200. 1600, and a network interface 1700.

The processor 1100 may be a central processing unit (CPU) or a semiconductor device that processes commands stored in the memory 1300 and/or the storage 1600 . The memory 1300 and the storage 1600 may include various types of volatile or nonvolatile storage media. For example, the memory 1300 may include read only memory (ROM) and random access memory (RAM).

Accordingly, the steps of a method or algorithm described in connection with the embodiments disclosed herein may be directly implemented as hardware executed by the processor 1100, a software module, or a combination of the two. A software module resides in a storage medium (i.e., memory 1300 and/or storage 1600) such as RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, or a CD-ROM. You may.

An exemplary storage medium is coupled to the processor 1100, and the processor 1100 can read information from, and write information to, the storage medium. Alternatively, the storage medium may be integral with the processor 1100. The processor and storage medium may reside within an application specific integrated circuit (ASIC). An ASIC may reside within a user terminal. Alternatively, the processor and storage medium may reside as separate components within a user terminal.

The above description is merely an example of the technical idea of the present invention, and various modifications and variations can be made to those skilled in the art without departing from the essential characteristics of the present invention.

Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention, but to explain, and the scope of the technical idea of the present invention is not limited by these embodiments. The protection scope of the present invention should be construed according to the claims below, and all technical ideas within the equivalent range should be construed as being included in the scope of the present invention.

100: device
200: voice recognition server
110: vehicle terminal
120: mobile communication terminal
111: audio signal input unit
112: communication department
113: display unit
114: storage unit
115: voice filtering unit
116: command filtering unit
117: verification unit
118: control unit

Claims (20)

a voice signal input unit that receives a voice signal including a wakeup keyword and a command;
a voice filtering unit that recognizes a wakeup keyword positioned at the head of the voice signal and wakes up a device corresponding to the wakeup keyword among a plurality of devices; and
a command filtering unit for selecting main commands based on words and verbs of commands of the voice signal;
including,
The command filtering unit,
If the command is one multi-command including a plurality of instructions, the number of verbs and the number of words of each of the plurality of instructions included in the one multi-command are determined, and the number of verbs or the number of words is the highest among the plurality of instructions. Voice recognition control device to select many commands. The method of claim 1,
Verification unit for verifying whether to transmit a vehicle command through the wake-up device
Voice recognition control device characterized in that it further comprises. The method of claim 2,
The verification unit,
If the wake-up device is at least one,
Checking whether the command of the voice signal matches the command grammar for each device, selecting a device with the highest grammar matching rate with the at least one wake-up device, and transmitting the vehicle command. Characterized in that Voice recognition control device. The method of claim 1,
The voice filtering unit,
Text is extracted from the voice signal, and if at least one wakeup keyword is present in the voice signal, voice signal correlation and text correlation are performed for the at least one wakeup keyword, and wake with the highest correlation is performed. A voice recognition control device characterized by selecting an up keyword. The method of claim 1,
The command filtering unit,
When a plurality of voice signals are input, a character matching rate and signal strength of each command are determined, and a voice signal having a high character matching rate or a high signal strength is selected. delete The method of claim 1,
The command filtering unit,
and selecting a command having the greatest number of verbs and words in the command, giving priority to the number of verbs over the number of words. The method of claim 1,
A biometric authentication unit that performs biometric authentication when the voice signal is input
Voice recognition control device characterized in that it further comprises. The method of claim 2,
When verification is completed by the verification unit, a feedback unit for feeding back that the wakeup keyword has been recognized to the user who uttered the voice signal
Voice recognition control device characterized in that it further comprises. The method of claim 1,
The voice filtering unit,
When the wakeup keyword exists in the voice signal, it is determined that the wakeup keyword is a valid wakeup keyword when the wakeup keyword is located at the beginning of the voice signal. For a plurality of devices having different wakeup keywords, recognizing a wakeup keyword located at the beginning of a user's voice signal, waking up a device corresponding to the recognized wakeup keyword, and filtering commands in the voice signal voice recognition control device; and
A voice recognition server for performing voice recognition on the voice signal received from the device and providing the voice recognition result to the device.
including,
The voice recognition control device,
If the command is one multi-command including a plurality of instructions, the number of verbs and the number of words of each of the plurality of instructions included in the one multi-command is determined, and the number of verbs or the number of words is the largest among the plurality of instructions. Voice recognition control system, characterized in that for selecting a command. The method of claim 11,
The device,
a voice signal input unit that receives a voice signal including a wakeup keyword and a command;
a voice filtering unit that recognizes a wakeup keyword positioned at the head of the voice signal and wakes up a device corresponding to the wakeup keyword among a plurality of devices;
a command filtering unit for selecting main commands based on words and verbs of commands of the voice signal; and
Verification unit for verifying whether to transmit a vehicle command through the wake-up device
Voice recognition control system comprising a. The method of claim 12,
The verification unit,
If the wake-up device is at least one,
Checking whether the command of the voice signal matches the command grammar for each device, selecting a device with the highest grammar matching rate with the at least one wake-up device, and transmitting the vehicle command. Characterized in that Voice recognition control system. The method of claim 12,
The voice filtering unit,
Text is extracted from the voice signal, and if at least one wakeup keyword exists in the voice signal, voice signal correlation and text correlation are performed for the at least one wakeup keyword, and the wakeup with the highest correlation is performed. Voice recognition control system characterized by selecting up keywords. The method of claim 12,
The command filtering unit,
When a plurality of voice signals are input, a character matching rate of each command and signal strength are determined, and a voice signal having a high character matching rate or a high signal strength is selected. delete Receiving a voice signal including a wakeup keyword and a command;
recognizing a wakeup keyword located at the head of the voice signal and waking up a device corresponding to the wakeup keyword among a plurality of devices;
selecting a main command based on words and verbs of the command of the voice signal;
verifying whether to transmit a vehicle command through the wake-up device; and
Transmitting vehicle commands through a verified device
including,
The step of selecting the command,
If the command is one multi-command including a plurality of instructions, the number of verbs and the number of words of each of the plurality of instructions included in the one multi-command are determined, and the number of verbs or the number of words is the highest among the plurality of instructions. A voice recognition control method characterized in that a number of commands are selected. The method of claim 17
The step of verifying whether to transmit the vehicle command is
Checking whether the command of the voice signal matches the command grammar for each device, selecting a device with the highest grammar matching rate with the at least one wake-up device and transmitting the vehicle command. Characterized in that Voice recognition control method. The method of claim 17
Waking up the device,
Text is extracted from the voice signal, and if at least one wakeup keyword exists in the voice signal, voice signal correlation and text correlation are performed for the at least one wakeup keyword, and the wakeup with the highest correlation is performed. A voice recognition control method comprising selecting an up keyword. The method of claim 17
The step of selecting the main command,
When a plurality of voice signals are input, a character matching rate and signal strength of each command are determined, and a voice signal having a high character matching rate or a large signal strength is selected.

Images