JP2008145676A - Speech recognition device and vehicle navigation device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To promptly perform speech recognition of a phrase which is desired to recognize by a user, while alleviating burden for uttering a phrase for a user. <P>SOLUTION: A speech recognition device 10 detects a position where the user holds a steering wheel 16, and changes a speech dictionary corresponding to the holding position in a plurality of voice dictionaries 15a to 15c, as a dictionary for use. When speech which the user utters is input, the speech uttered by the user is recognized by collating a speech waveform of the input speech and a similar waveform of the phrase registered in the speech dictionary which is changed as the dictionary for use. The user does not need to utter the phrase for changing one of the dictionaries as the dictionary for use, in the plurality of speech dictionaries 15a to 15c, and the user may only utter the phrase which is desired to recognize. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention recognizes a voice uttered by a user by collating a voice waveform of a voice uttered by a user with an approximate waveform of a word registered in a voice dictionary switched from a plurality of voice dictionaries as a use dictionary. The present invention relates to a voice recognition device and a vehicle navigation device.

For example, in a speech recognition device mounted on a vehicle navigation device, a speech waveform of a user uttered and an approximate waveform of a phrase registered in a speech dictionary are collated to recognize the speech uttered by the user. (For example, refer to Patent Document 1).
JP 2005-121966 A

  By the way, in recent years, with the addition of functions, voice recognition has been used in a wide variety of situations. Along with this, for example, a voice dictionary for addresses, a voice dictionary for facility names, and navigation function control Prepare multiple voice dictionaries according to the purpose of the voice dictionary, etc., first switch one of the multiple voice dictionaries as the use dictionary, then switch as the voice waveform of the voice you want to recognize and the use dictionary There is provided a configuration for recognizing a speech to be recognized by collating with an approximate waveform of a word registered in the speech dictionary.

  In this case, the user first utters a phrase for switching one of the plurality of voice dictionaries as a use dictionary, and then utters a phrase to be voice-recognized. Specifically, the user can first switch the address voice dictionary as a use dictionary by, for example, saying “ji” “yu” “u” “shi” “yo”. By saying “a”, “b”, “chi”, “ke”, and “n”, the word “Aichi” that is desired to be recognized can be recognized.

  However, in this case, since two words, a word for switching the dictionary to be used and a word to be recognized by voice, are recognized by voice, it is troublesome for the user to utter the word twice. There is a problem that it takes time until speech recognition is performed for a phrase to be recognized, which is inferior in operability.

  The present invention has been made in view of the above-described circumstances, and an object of the present invention is to reduce the burden on a user to speak a phrase and to quickly recognize a phrase that the user wants to recognize by speech. It is possible to provide a voice recognition device and a vehicle navigation device that can improve operability.

  According to the first aspect of the present invention, the operation position detecting means detects a user's operation position from among a plurality of operation positions set as operation targets corresponding to a plurality of voice dictionaries, and switches voice dictionaries. The means switches the voice dictionary corresponding to the operation position detected by the operation position detection means from among a plurality of voice dictionaries as a use dictionary. When the voice input unit inputs the voice uttered by the user, the voice recognition unit and the speech waveform input by the voice input unit and the phrase registered in the voice dictionary switched by the voice dictionary switching unit The voice generated by the user is recognized by collating with the approximate waveform.

  Accordingly, it is necessary for the user to utter a phrase for switching any one of the plurality of voice dictionaries as the use dictionary by switching any one of the plurality of voice dictionaries based on the operation position of the user. It is sufficient that the user speaks only the words and phrases that the user wants to recognize, and the burden on the users to speak the words can be reduced, and the user can quickly recognize the words and phrases that the user wants to recognize. Can increase the sex.

  According to the second aspect of the present invention, the voice dictionary switching unit is configured to provide a voice corresponding to the operation position detected by the operation position detection unit from among a plurality of voice dictionaries before the own device shifts to the voice recognition standby state. Since the dictionary is switched as the use dictionary, when the device shifts to the voice recognition standby state, one of the plurality of voice dictionaries can be already switched as the use dictionary, and the word or phrase that the user wants to recognize by voice can be selected. Speech can be recognized more promptly.

  According to the third aspect of the present invention, the operation target is steering, the operation position detection unit detects a position where the user is grasping the steering, and the voice dictionary switching unit is selected from the plurality of voice dictionaries. Since the speech dictionary corresponding to the steering wheel position detected by the operation position detection means is switched as the use dictionary, any one of the plurality of speech dictionaries can be switched as the use dictionary by switching the position at which the user grips the steering wheel. .

  According to the invention described in claim 4, since the operation target is a part of the steering wheel, one of the plurality of voice dictionaries is selected by grasping the portion that is the steering operation target with one hand. It can be switched as a use dictionary, and it can be operated by grasping a part that is not subject to steering operation with the other hand, and any of a plurality of voice dictionaries can be used without causing any trouble in driving operation. Can be easily switched as a use dictionary.

  According to the invention described in claim 5, the grip pressure detection means detects the grip pressure at which the user is gripping the steering wheel, and if the grip pressure detected by the grip pressure detection means is equal to or greater than a threshold value, Transitions to the voice recognition standby state, so that the user transitions to the voice recognition standby state by holding the steering wheel with a gripping pressure equal to or higher than the threshold value, for example, without operating the other switch by releasing the hand from the steering wheel. Therefore, the operability can be further enhanced.

  According to the invention described in claim 6, since the voice dictionary assignment switching means switches the assignment of the voice dictionary to the steering position of the steering wheel, for example, the voice dictionary having a high frequency of use is assigned to a position that is easy for the user to grip in the steering wheel. As a result, the operability can be further enhanced.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 2 shows the configuration of the vehicle navigation apparatus as a functional block diagram. The vehicle navigation device 1 includes a control device 2, a position detector 3, a map data storage device 4, an operation switch group 5, a vehicle signal input / output device 6, a communication device 7, a VICS receiver 8, a voice control device 9, and a voice recognition device. 10, a display device 11, a memory 12, and a remote control sensor 13.

  The control device 2 includes a CPU, a ROM, a RAM, an I / O interface, a bus for connecting them (not shown), and controls the overall operation of the vehicle navigation device 1. The position detector 3 includes a G sensor 3a, a gyroscope 3b, a distance sensor 3c, and a GPS receiver 3d, and each component of the position detector 3 has detection errors having different properties. In this case, when a detection signal is input from each component of the position detector 3, the control device 2 complements the input detection signals to detect (specify) the current position and traveling direction of the host vehicle.

  The map data storage device 4 stores map data transferred from a recording medium 14 such as a DVD-ROM. In this case, the recording medium 14 records the map data used for route guidance, and as shown in FIG. 1, as the voice dictionary, the address voice dictionary 15a, the facility name voice dictionary 15b, and the navigation function control. A plurality of (three in the present embodiment) voice dictionaries corresponding to the application such as the voice dictionary 15c for use are recorded. The recording medium 14 may be, for example, an HDD or a memory card. The operation switch group 5 includes a mechanical switch disposed around the display device 11, a touch switch formed on, for example, a color liquid crystal display of the display device 11, a PTT switch disposed at a predetermined position in the vehicle interior, and the like. Has been.

  The vehicle signal input / output device 6 inputs / outputs various signals to / from various sensors and in-vehicle devices mounted on the vehicle. For example, the vehicle signal input / output device 6 inputs a vehicle speed signal indicating the vehicle speed from the vehicle speed sensor, and the steering 16 (in the present invention). A grip position signal indicating the position (grip position) where the user is gripping the steering wheel 16 is input. As shown in FIG. 3, the steering 16 has three pressure sensors 17a to 17c having an evenly embedded shape embedded in the left half region of the annular rim portion when viewed from the front. The pressure sensors 17a to 17c indicate grip positions indicating that the user is gripping the corresponding portion of the steering wheel 16 when a pressure of a predetermined value or more is detected in response to the user gripping the corresponding portion of the steering wheel 16. Output a signal. In this case, each of the pressure sensors 17a to 17c includes a numerical value indicating a grip pressure at which the user is gripping the steering wheel in the grip position signal and outputs the grip position signal.

  Here, the locations where the respective pressure sensors 17a to 17c are embedded correspond to the three voice dictionaries 15a to 15c recorded on the recording medium 14 described above. Specifically, the location where the pressure sensor 17a is embedded corresponds to the address speech dictionary 15a, and the location where the pressure sensor 17b is embedded corresponds to the facility name speech dictionary 17b. The location where the pressure sensor 17c is embedded corresponds to the voice dictionary 17c for navigation function control.

  The communication device 7 communicates with a mobile communication network. The VICS receiver 8 receives VICS information from the outside. The sound control device 9 performs sound control on the sound input by the microphone 18 disposed at a predetermined position in the vehicle interior and the sound output by the speaker 19 disposed at a predetermined position in the vehicle interior.

  As shown in FIG. 1, the speech recognition apparatus 10 includes a grip position detection unit 20a (operation position detection unit and grip pressure detection unit in the present invention) and a voice dictionary switching unit 20b (speech dictionary in the present invention) for each function. Switching means, voice dictionary assignment switching means), voice dictionary storage section 20c, voice input section 20d (speech input means in the present invention), voice recognition section 20e (speech recognition means in the present invention) and recognition result output section 20f. It is prepared for.

  When the grip position detection unit 20a inputs the grip position signal output from the steering 16 via the vehicle signal input / output device 6, the grip position signal is analyzed to analyze the position at which the user is gripping the steering wheel 16 (grip position). And the grip pressure at which the user is gripping the steering wheel is also detected. The voice dictionary switching unit 20b switches (selects) one of the address voice dictionary 15a, the facility name voice dictionary 15b, and the navigation function control voice dictionary 15c recorded on the DVD 14 as a use dictionary. Further, for example, when the user performs an operation of switching the assignment of the voice dictionary with the operation switch group 5, the voice dictionary switching unit 20 b is recorded on the recording medium 14 and the place where the above-described pressure sensors 17 a to 17 c are embedded. The correspondence (allocation) with the three voice dictionaries 15a to 15c is switched. The speech dictionary storage unit 20c stores the speech dictionary switched by the speech dictionary switching unit 20b as a use dictionary.

  The voice input unit 20 d inputs the voice input by the microphone 18 via the voice control device 9. When the voice input unit 20d inputs a voice, the voice recognition unit 20e refers to the voice dictionary stored in the voice dictionary switching unit 20b, that is, the voice dictionary switched by the voice dictionary switching unit 20b as a use dictionary. The speech waveform input by the unit 20d is compared with the approximate waveform of a phrase registered in the speech dictionary stored in the speech dictionary storage unit 20c as a use dictionary, so that speech uttered by the user is recognized as speech. The recognition result output unit 20f outputs the recognition result recognized by the voice recognition unit 20e to the control device 2.

  Based on the recognition result input from the speech recognition device 10 in this manner, the control device 2 searches for a destination or changes the scale of the map display, for example. Note that the voice recognition device 10 is in a voice recognition standby state, for example, while the PTT switch is pressed, and can perform the voice recognition described above.

  The display device 11 is composed of, for example, a color liquid crystal display, and displays a current position graphic representing the current position of the host vehicle and a travel locus on a map corresponding to map data. The display device 10 may be composed of an organic EL, a plasma display, or the like. The memory 12 is composed of, for example, a removable flash memory card. The remote control sensor 13 receives the operation signal transmitted from the operation remote controller 21 and outputs it to the control device 2.

Next, as an operation of the above-described configuration, processing performed by the speech recognition apparatus 10 will be described with reference to a flowchart shown in FIG.
When the voice recognition device 10 shifts to a voice recognition standby state in response to the user pressing the PTT switch, the voice recognition device 10 analyzes the grip position signal input from the steering 16 via the vehicle signal input / output device 6 at that time. Then, the position where the user is gripping the steering wheel 16 (grip position) is detected by the grip position detection unit 20a (step S1).

  Next, the speech recognition apparatus 10 refers to the speech dictionary stored in the speech dictionary storage unit 20c at that time, that is, the speech dictionary switched by the speech dictionary switching unit 20b as a use dictionary (step S2). It is determined whether or not the voice dictionary that is switched as the use dictionary by the switching unit 20b corresponds to the detected grip position (step S3).

Here, when the voice recognition device 10 determines that the voice dictionary switched by the voice dictionary switching unit 20b as the use dictionary corresponds to the detected grip position ("YES" in step S3), the user recognizes the voice dictionary. While waiting for input of the uttered voice (step S6), it is determined whether or not the voice recognition waiting state has been released (step S7). If the voice recognition device 10 determines that the voice uttered by the user has been input by the voice input unit 20d ("YES" in step S6), the voice waveform of the input voice and the voice switched as the use dictionary are determined. The speech uttered by the user by collating with the approximate waveform of the word registered in the dictionary is recognized by the speech recognition unit 20e (step S8), and the recognition result of the speech recognition is output to the control device 2 (step S8). S9),
Returning to step S1, the above-described series of processing is repeated.

  On the other hand, when the speech recognition device 10 determines that the speech dictionary switched as the use dictionary by the speech dictionary switching unit 20b does not correspond to the detected grip position (“NO” in step S3), the detected grip The speech dictionary corresponding to the position is switched as a use dictionary (step S4), and the speech dictionary switched as the use dictionary is stored in the speech dictionary storage unit 20c (step S5). Then, the voice recognition device 10 waits for input of the voice uttered by the user (step S6), determines whether or not the voice recognition waiting state is canceled (step S7), and inputs the voice uttered by the user. If it is determined ("YES" in step S6), then steps S8 and S9 described above are performed.

  When the voice recognition device 10 determines that the voice recognition standby state has been released in response to the release of the PTT switch being released (“YES” in step S7), the above-described series of processing ends.

  By the way, the above has described the configuration in which the user shifts to the voice recognition waiting state in response to the user pressing the PTT switch. However, the voice recognition is performed on condition that the grip pressure detected by the grip position detection unit 20a is equal to or greater than a threshold value. It may be configured to shift to a waiting state. In addition, the configuration in which the voice dictionary is switched according to the position where the user is holding the steering wheel 16 after shifting to the voice recognition waiting state in response to the user pressing the PTT switch has been described. Even if the user is not pressing the PTT switch when activated, the voice dictionary may be switched according to the position where the user is holding the steering wheel 16.

  As described above, according to the present embodiment, the voice recognition device 10 detects the position where the user is gripping the steering wheel 16, and selects a voice dictionary corresponding to the grip position from the plurality of voice dictionaries 15 a to 15 c. When a voice uttered by a user is input after switching as a use dictionary, the voice uttered by the user is checked by comparing the voice waveform of the input voice with the approximate waveform of a phrase registered in the voice dictionary switched as the use dictionary. Since the speech recognition is configured, it is not necessary for the user to utter a word or phrase for switching any one of the plurality of speech dictionaries 15a to 15c as the use dictionary, and only the word or phrase that the user wants to recognize by speech is spoken. It is possible to reduce the burden of speaking a phrase for the user, and quickly recognize the phrase that the user wants to recognize. It can, it is possible to improve the operability.

  Further, if the speech recognition device 10 is configured to switch the speech dictionary corresponding to the grip position as the use dictionary from among the plurality of speech dictionaries 15a to 15c before the speech recognition device 10 shifts to the speech recognition standby state, the speech recognition device 10 can be used. At the time of shifting to the voice recognition standby state, any one of the plurality of voice dictionaries 15a to 15c can be already switched as the use dictionary, and the user can more quickly recognize the words that the user wants to recognize by voice. it can.

  In addition, since the position where the user is grasping the steering wheel 16 is detected using the left half portion of the steering wheel 16, the left half portion of the steering wheel 16 is grasped with the left hand so that a plurality of voice dictionaries 15 a to 15 c can be stored. One of them can be switched as a use dictionary, and a driving operation can be performed by grasping the right half of the steering wheel 16 with the right hand. Among the plurality of voice dictionaries 15a to 15c, there is no hindrance to the driving operation. Can be easily switched as a use dictionary.

  In addition, when the grip position detection unit 20a is configured to shift to the voice recognition waiting condition on condition that the grip pressure detected by the grip position detection unit 20a is equal to or greater than the threshold, the user releases the hand from the steering wheel 16 and operates another switch, for example. Even if the user does not hold the steering wheel 16 and grips it with a gripping pressure equal to or higher than the threshold, the user can shift to a voice recognition standby state and can further improve operability.

  Furthermore, for example, when the user performs an operation of switching the assignment of the voice dictionary with the operation switch group 5, the assignment of the voice dictionary to the grip position of the steering wheel 16 is switched. By assigning a voice dictionary having a high use frequency to, operability can be further improved.

The present invention is not limited to the above-described embodiment, and can be modified or expanded as follows.
The operation target is not limited to the steering, but may be another device (for example, a shift lever) that can be operated by the user during driving within a range in which the driving operation is not hindered.
The means for detecting the position where the user is grasping the steering is not limited to the configuration in which the pressure sensor is used, and may be a configuration in which another sensor such as an electrode is used. Moreover, the structure by which a sensor is arrange | positioned at the right half part, lower part, or upper part of steering may be sufficient.
The voice dictionary may be used for purposes other than the address voice dictionary, the facility name voice dictionary, and the navigation function control voice dictionary.
A configuration in which the voice recognition device shifts to a voice recognition standby state triggered by a user operation other than the user pressing the PTT switch or the user gripping the steering wheel with a gripping pressure equal to or higher than a threshold value may be used.

1 is a functional block diagram of a speech recognition apparatus according to an embodiment of the present invention. Functional block diagram of vehicle navigation device The figure which shows correspondence with the grip position of the steering wheel and the voice dictionary flowchart

Explanation of symbols

  In the drawings, 1 is a vehicle navigation device, 10 is a speech recognition device, 15a to 15c are speech dictionaries, 16 is a steering (operation target), 20a is a grip position detection unit (operation position detection means, grip pressure detection means), and 20b is A speech dictionary switching unit (speech dictionary switching unit, speech dictionary allocation switching unit), 20d is a speech input unit (speech input unit), and 20e is a speech recognition unit (speech recognition unit).

Claims (7)

A voice dictionary switching means for switching one of a plurality of voice dictionaries as a use dictionary; a voice input means for inputting a voice uttered by a user; a voice waveform of the voice inputted by the voice input means; and the voice dictionary switching means. A speech recognition device comprising speech recognition means for recognizing speech uttered by the user by collating with an approximate waveform registered in a speech dictionary switched as a use dictionary,
An operation position detecting means for detecting a user's operation position from among a plurality of operation positions set as operation targets corresponding to a plurality of voice dictionaries;
The voice recognition apparatus characterized in that the voice dictionary switching means switches a voice dictionary corresponding to the operation position detected by the operation position detection means from a plurality of voice dictionaries as a use dictionary. The speech recognition apparatus according to claim 1,
The voice dictionary switching means switches the voice dictionary corresponding to the operation position detected by the operation position detection means from among a plurality of voice dictionaries as a use dictionary before the apparatus shifts to a voice recognition standby state. Voice recognition device. The speech recognition apparatus according to claim 1 or 2,
The operation target is steering,
The operation position detecting means detects a position where the user is holding the steering wheel,
The speech recognition apparatus characterized in that the speech dictionary switching means switches a speech dictionary corresponding to a steering grip position detected by the operation position detection means from a plurality of speech dictionaries as a use dictionary. The speech recognition apparatus according to claim 3,
The speech recognition apparatus characterized in that the operation target is a part of a steering wheel. The speech recognition device according to claim 3 or 4,
A grip pressure detecting means for detecting a grip pressure at which the user is gripping the steering wheel;
The speech recognition apparatus according to claim 1, wherein when the grip pressure detected by the grip pressure detection means is equal to or greater than a threshold value, the apparatus shifts to a speech recognition standby state. The speech recognition device according to any one of claims 3 to 5,
A speech recognition apparatus comprising speech dictionary assignment switching means for switching assignment of a speech dictionary to a steering position of a steering wheel. A voice recognition device according to any one of claims 1 to 6, comprising:
A vehicle navigation device that performs a process related to navigation based on a recognition result recognized by the voice recognition device.

Images