JP5014662B2 - On-vehicle speech recognition apparatus and speech recognition method - Google Patents

Description

  The present invention relates to a technology for controlling an in-vehicle device or the like using a speech recognition function, and more particularly to an in-vehicle speech recognition apparatus and speech recognition adapted to perform speech recognition corresponding to an operation instruction spoken by an occupant in a vehicle interior. Regarding the method.

  Recent vehicles are equipped with devices and devices for providing various services to drivers, passengers in the passenger seat, passengers in the rear seat, etc. (hereinafter also referred to as “user” for convenience). . Typical in-vehicle devices include a navigation device equipped with a function (route guidance function) for guiding the user to travel to a set destination without making a mistake, and various sources (radio receiver, CD player, TV) There are audio / video (A / V) devices that provide various entertainment such as audio (audio) information and video (video) information output from a receiver, a DVD player, etc., and an air conditioner. These in-vehicle devices (apparatuses) are operated according to the operation instructions when the user gives a necessary instruction by operating the remote controller or the operation panel. The changed operating state of the device can be heard via a speaker (such as wireless headphones for a user at the rear seat) installed in the passenger compartment, or can be viewed through a screen of a display device such as an in-vehicle monitor.

  As described above, a required operation instruction can be input to each in-vehicle device based on a manual operation such as a remote control operation. However, recently, the control of the device can be performed only by voice input (speech) of the operation instruction. Devices equipped with a function that can be performed (voice recognition function) have also appeared. Such a voice recognition function is advantageous in terms of convenience for the user's operation, and is particularly useful for the driver in terms of safe driving.

  In order to realize this voice recognition function, a voice recognition dictionary is required. In this voice recognition dictionary, words such as words and phrases to be voice recognition, that is, in-vehicle devices to be controlled based on voice recognition ( Hereinafter, the vocabulary related to the operation instruction of “control target device” is registered in advance. For example, a vocabulary such as “destination”, “menu”, “surrounding search” is registered for a navigation device, and “radio”, “FM”, “AM”, “menu” for an A / V device. Vocabulary such as “,” “play” and “stop” is registered.

As a technique related to the above-described conventional technique, for example, as described in Patent Document 1, in a voice control apparatus that performs speech recognition of content uttered by a user and controls the control target equipment, the operation of the control target equipment There is a device that can appropriately perform a voice operation of a control target device by recognizing a user's utterance in consideration of a state.
JP 2004-86150 A

  As described above, in the conventional technology, a function that can control the device only by uttering an operation instruction to the device to be controlled is realized, but in the conventional method, all the vocabularies registered in the speech recognition dictionary are stored. On the other hand, since speech recognition is performed, if the number of registered vocabularies increases, inconveniences described below may occur.

  That is, the speech recognition engine calculates the degree of match between the content (speech command) uttered by the user and all vocabularies (commands) registered in the speech recognition dictionary, and the command with the highest degree of match is calculated from the calculation result. It is determined as a voice command issued by the user (voice recognition). At this time, there is no problem as long as the command with the highest degree of matching can be identified, but as the number of registered vocabularies increases, the number of words that are similar to “reading” in pronunciation also increases, so voice recognition In the engine, it is not always possible to specify one, and as a result, a vocabulary that does not match may be erroneously recognized. In other words, in the conventional speech recognition method, when the number of vocabularies registered in the speech recognition dictionary to be used increases, the proportion of erroneous recognition increases accordingly, and the user's utterance content is recognized accurately. Is difficult (recognition rate for voice commands decreases).

  The present invention has been created in view of the problems in the prior art, and when recognizing speech content to control an in-vehicle device, the on-vehicle speech recognition device and speech recognition capable of increasing the recognition rate for the speech content. It aims to provide a method.

In order to solve the above-described problems of the prior art, according to one aspect of the present invention, voice input means for voice input of information instructed by a user in a vehicle interior and a user who speaks via the voice input means are specified. Stores a speech recognition dictionary in which vocabulary related to operation instructions of each of a plurality of control target devices is registered, and a vocabulary to be recognized for each control target device is set in advance. Stipulates the relationship between the dictionary storage means storing the first table that defines the relationship with the weighting, the seat on which the user is seated, and the control target device that is the source of information viewed by the user of the seat memory means for storing a second table, the voice input unit, speaker identifying means, and a operatively connected to the control means in the dictionary storing means and memory means, said control hand , Upon identifying the speaker identification means cooperating with speaker, the speech recognition dictionary and the first, with reference to the second each table, the source of information which the speaker is viewing A predetermined weight is added to the vocabulary corresponding to the control target device, and the speech content of the speaker is recognized by referring to the weighted vocabulary, and control according to the recognized speech content is performed. An in-vehicle speech recognition apparatus is provided that is performed on the control target device.

According to the in-vehicle speech recognition device according to this aspect , when a user (speaker) who has spoken in the passenger compartment is specified, the second table is referred to as a source of information viewed by the speaker. identify the control target device, among vocabularies registered in the voice recognition dictionary which is shared to the source (control target device), vocabulary only corresponding to the specified control target device, is defined in the first table The predetermined weighting is added, and speech recognition is performed on the utterance content of the speaker by referring to the vocabulary to which the weighting is added.

As a result, when recognizing the content spoken by the user, among the vocabulary registered in the speech recognition dictionary, the vocabulary to which weighting is added by referring to the first table (refer to the second table). The vocabulary that does not match is compared with the case where speech recognition is performed for all vocabularies registered in the speech recognition dictionary as in the past. The rate of misrecognition can be reduced. That is, the recognition rate for the content (voice command) spoken by the user can be improved.

According to another aspect of the present invention, in an in-vehicle speech recognition apparatus having a function of performing speech recognition corresponding to an operation instruction for a control target device spoken by a user in a vehicle interior, a plurality of control target devices are preliminarily provided. A first table that defines the relationship between a vocabulary to be recognized for each control target device and a preset weight, together with one speech recognition dictionary in which vocabulary related to each operation instruction is registered, and a user seated A second table that defines the relationship between the seat being seated and the control target device that is the source of information viewed by the user of the seat is stored in the storage means, and the utterance is detected when the utterance is detected. identify who the voice recognition dictionary, and the first, with reference to the second each table, a predetermined vocabulary to which the speaker is corresponding to the control target device is the source of the information being viewed Adding weighting, referring to the weighted vocabulary, performing speech recognition on the utterance content of the speaker, and performing control according to the recognized utterance content on the control target device A speech recognition method is provided.

  Other structural features of the in-vehicle speech recognition apparatus and speech recognition method according to the present invention and specific processing modes based on the features will be described in detail with reference to embodiments of the invention to be described later.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

  FIG. 1 shows a configuration of an in-vehicle audio / video (A / V) navigation system incorporating an in-vehicle voice recognition device according to an embodiment of the present invention.

  As shown in the figure, the in-vehicle A / V / navigation system 40 includes an in-vehicle voice recognition device AR that characterizes the present invention, and a target device in which control corresponding to the utterance content (voice command) is performed based on the voice recognition result. (In the example shown, the radio receiver 1, the CD player 2, the DVD player 3, the TV receiver 4, the navigation unit 5, and the air conditioner 6) and the user at the front seat perform various setting operations on the devices to be controlled. Operation unit for the front seat (head unit (H / U)) 20 and a rear seat operation unit for the rear seat user to perform various setting operations on each control target device (excluding the navigation unit 5) 30, a front seat display unit 25, an amplifier unit 26, a speaker 27, a rear seat display unit 31, and a wireless head. It is constituted by a phone 32. The in-vehicle voice recognition device AR, each source (device to be controlled) 1 to 6, the front seat operation unit 20, each display unit 25, 31 and the amplifier unit 26 are connected to a bus 7 such as an optical fiber provided as a transmission path. Are connected to each other.

  In the example shown in the figure, only one speaker 27 is shown, but in reality, the required number in a predetermined position in the passenger compartment, for example, at least the vicinity of the left and right of the rear seats when the rear seats are one row. In total, four speakers 27 are installed in the vicinity of the left and right sides of the front seat. Similarly, only one unit (one) is shown for the operation unit 30 for the rear seat, the display unit 31 and the wireless headphones 32, but in reality, the required number according to the number of passengers in the rear seat, For example, if the rear seats are in a single row, two (two) operation units 30, a display unit 31, and wireless headphones 32 are provided for the left and right passengers of the rear seats.

  The in-vehicle voice recognition device AR that characterizes the present invention includes a hard disk drive (HDD) 8 as a storage medium, a microphone array 9, and a voice recognition unit 10. Storage areas to which data (speech recognition dictionary) used when executing the speech recognition function are allocated to the disk (not shown) driven by the HDD 8 together with map data used when the navigation function is executed. Stored in The map data is divided into longitude and latitude widths of appropriate sizes according to each scale level (1/12500, 1/25000, 1 / 50,000, etc.), and can be used for various processes such as route search and map matching. Necessary road unit data, intersection unit data showing details of the intersection, data on various facilities (convenience store, gas station, super discount shop, etc.) (various information such as location, address, telephone number, genre, etc.) Contains. The contents of the speech recognition dictionary stored in the HDD 8 will be described later.

  The microphone array 9 is configured by arranging a plurality of microphones in an array at predetermined intervals, and is appropriately installed, for example, in the vicinity of a sun visor or a room mirror in front of the driver's seat in the passenger compartment. The microphone array 9 (each microphone) detects an instruction (speech) related to the operation of the control target device spoken by the user (driver, passenger in the passenger seat, or rear passenger) according to the sound pressure level. This is converted to an analog audio signal. As will be described later, the signal detected by each microphone is used to specify the location of the user (speaker) who speaks in the passenger compartment, that is, the speaker. This speaker specifying method will be described later together with the internal configuration of the speech recognition unit 10.

  The front seat operation unit (H / U) 20 is installed in the form of an “operation panel” on the center console between the two seats so that the driver and the passenger in the front passenger seat can share the display. The unit 25 is disposed above the operation panel (H / U). The display unit 25 includes, for example, a dual display type LCD monitor (referred to as “dual display” for the sake of convenience). This is the same as when the same screen is viewed from the right (driver's side) and left (passenger seat). Different images can be displayed at the same time when viewed from the side. On the screen of the dual display (display unit 25), various video information output from the navigation unit 5 (map around the vehicle position, guidance route from the vehicle position to the destination, facility based on voice recognition) Information such as search information), video information output from a video source such as the DVD player 3 or the TV receiver 4, and the like are displayed.

  On the other hand, the rear seat operation unit 30 is provided in the form of a “remote control” so that the user at the rear seat can easily operate, and is connected to the corresponding rear seat display unit 31 by infrared communication. The rear seat display unit 31 is installed, for example, in the rear part of the headrest of the front seat, and has an LCD monitor or the like that displays video information on the display screen, like the front display unit 25. The display unit 31 is connected to the corresponding wireless headphones 32 by infrared communication and RF communication.

  Each source (device to be controlled) 1 to 6 has, as a basic operation, data related to an operation instruction sent from the front seat operation unit 20 to the bus 7 or infrared communication from the rear seat operation unit (remote control) 30. Receives the data related to the operation instruction sent to the bus 7 via the display unit 31 or the data related to the operation instruction sent to the bus 7 from the voice recognition unit 10 ("device control signal" described later), Based on the data related to the operation instruction, its own operation state is set or changed, and data indicating the result (current operation state) is sent to the bus 7 as an audio / video signal. For example, in the case of the radio receiver 1, in response to an operation instruction given from each operation unit 20, 30 or the voice recognition unit 10, an audio signal is generated by receiving and demodulating an FM broadcast or AM broadcast signal. This is converted into digital audio data and sent to the bus 7. In the case of the DVD player 3, in response to an operation instruction given in the same manner, a signal recorded on the recording surface of the DVD selected by the user is read and the reproduced video data is sent to the bus 7.

  The front seat operation unit 20 includes a control unit 21, an operation unit 22, a display unit 23, and a memory unit 24. Among these, the operation unit 22 includes operation keys for performing various setting operations on the sources (control target devices) 1 to 6, for example, a power key for power on / off and volume adjustment, and each source. Select key for selecting, numeric key, preset key for performing a predetermined function, shift key with an arrow (such as FF / REW operation, seek up / down operation by operating the arrow part) Instructing the operation). The display unit 23 is arranged in the form of an LCD or the like on the operation panel (H / U), and based on data output from the control unit 21, various information, for example, FM / AM for the radio receiver 1 is used. For the CD player 2, the disc number and playback position (number of tracks, elapsed time, etc.) during the CD performance are displayed.

  The memory unit 24 is composed of a nonvolatile semiconductor memory such as a flash memory, and stores necessary information (data) based on control from the control unit 21. The memory unit 24 stores the device at the time when the output operation of the audio / video signal from the selected source (control target device) is stopped based on the operation instruction given from each operation unit 20, 30 or the voice recognition unit 10. The data indicating the operation state (hereinafter referred to as “apparatus operation state data”) is stored. This equipment operation state data is stored for reference when necessary at the start of the next output operation. The device operation status data includes, for example, a “source type” that indicates which device (source) was used, and an adjustment value of the volume and sound quality when listening to the sound of an audio source. “Volume / sound quality” to be instructed, “detailed information by source” to instruct a detailed device operation state for each source, and the like are included. The source-specific detailed information includes, for example, FM / AM type and broadcast station (reception frequency) information when the radio receiver 1 is used, and when the CD player 2 is used. Includes the disc number indicating which CD was being played among multiple CDs loaded, and information on the playback position indicating how much time had elapsed since the beginning of what song. It is.

  The control unit 21 is configured by a microcomputer or the like, and controls the entire system 40. Basically, audio / video data sent from the selected source (control target device) via the bus 7 is acquired based on the operation instructions given from the operation units 20 and 30 or the voice recognition unit 10. Control is performed such as an operation for reproducing audio / video information, an operation for displaying information instructing an operation state and an operation state on the display unit 23, a storage operation and a read operation for device operation state data. In this case, the acquired audio data is sent to the amplifier unit 26 via the bus 7 by the control unit 21 and appropriately D / A converted, and the volume and sound quality are controlled and amplified. 27 is output as audio. The acquired video data is sent to the display unit 25 by the control unit 21 via the bus 7 and displayed on the display screen as video information.

  On the other hand, the rear seat operation unit (remote control) 30 is not particularly shown, but an operation unit having the same function as the operation unit 22 on the front side and a signal corresponding to an operation instruction input from the operation unit are infrared rays. An infrared transmission unit for transmitting to the display unit 31 by communication. The rear seat display unit 31 is not particularly shown, but is equivalent to the infrared communication unit for communicating control signals and data between the remote controller 30 and the wireless headphones 32, and the front side control unit 21. A control unit that performs control, a display unit that includes an LCD monitor or the like similar to the front-side display unit 25, and a memory unit similar to the front-side memory unit 24 are provided.

<First Embodiment (see FIGS. 2 to 4)>
FIG. 2 schematically shows a part of the configuration of the in-vehicle speech recognition apparatus according to the first embodiment.

  The in-vehicle speech recognition apparatus AR according to the present embodiment includes an HDD 8, a microphone array 9, a digital signal processor (DSP) 11, a CPU 12, and a memory unit 13 including a RAM and the like as illustrated. Among these, the DSP 11, the CPU 12, and the memory unit 13 constitute the voice recognition unit 10 (FIG. 1). The DSP 11 includes a voice input unit 11a, a beam forming unit 11b, and a sound source direction specifying unit 11c as functional blocks. On the other hand, the CPU 12 includes a recognition dictionary selection unit 12a, a speech recognition processing unit 12b, and a device control signal generation unit 12c as functional blocks.

  The HDD 8 has a dedicated vocabulary associated with the operation instruction of the device corresponding to the source (control target device) to be controlled based on the voice recognition (that is, a vocabulary frequently issued as a voice command to the device). ) Are stored in advance. In the illustrated example, the vocabulary related to the navigation unit 5 (such as “destination”, “menu”, “surrounding search”, “current location”) and the vocabulary related to the DVD player 3 (“ Menu "," play "," stop ", etc.) dedicated recognition dictionary D2 and dedicated recognition dictionary D3 registered vocabulary (" radio "," FM "," AM ", etc.) related to the radio receiver 1 These three types of dictionaries are stored.

  The memory unit 13 includes information (management table) indicating a relationship between a seat (driver's seat, front passenger seat, rear seat) where the user is seated and a source (control target device) of information viewed by the user of the seat. ) Is stored. This management table is stored in the memory unit 24 in each unit in cooperation with the CPU 12, the control unit 21 in the front seat operation unit 20, and the control unit (not shown) in the rear seat display unit 31. It is created based on the “apparatus operating state data”. Therefore, when the operating state of each source (control target device) is changed, the contents of the management table are updated accordingly.

  In this embodiment, the direction of the sound source (in this case, the direction of the seat on which the speaker is seated) is specified using the microphone array 9 and the DSP 11 that processes the detection signal. A method for specifying the direction of a sound source using a plurality of microphones is known. In other words, each microphone is omnidirectional, but a plurality of microphones are arranged in an array so that sound from a sound source is detected by each microphone, and each detected data is added and processed to add directionality. it can. For example, as shown in FIG. 3, when sound enters from directly in front of the microphone array 9 (in the illustrated example, when a user P3 in the rear seat is speaking), the sound pressure reaching each microphone of the microphone array 9 Since the signals are almost in phase in phase, adding these results in a large level signal. On the other hand, when the sound is incident obliquely (in the example shown, when the driver seat user P1 and the passenger seat user P2 are speaking), a difference occurs in the time to reach each microphone, and in phase Since they are shifted in the positive direction or the negative direction, if these are added, they cancel each other and become a small signal in terms of level. Using this principle, it is possible to specify the direction of sound arrival (that is, the direction in which the speaker is present) based on the level and phase difference of the signal detected by each microphone. In the identification, the beam forming method is used in the present embodiment.

  That is, in the voice recognition unit 10, signals (analog voice signals) detected by each microphone of the microphone array 9 are appropriately amplified through the voice input unit 11a of the DSP 11, digitized, and then input to the beam forming unit 11b. And input to the voice recognition processing unit 12b of the CPU 12. The beam forming unit 11b performs direction estimation based on the input signal to generate a beam signal (beam forming), and based on the generated beam signal, the sound source direction specifying unit 11c generates a signal having a high sound pressure level. The receiving direction is specified as the direction of the sound source (the direction in which the speaker is present).

  The CPU 12 refers to the management table stored in the memory unit 13 by the recognition dictionary selection unit 12a and selects the identified speaker from the plurality of dedicated recognition dictionaries D1 to D3 stored in the HDD 8. A dedicated speech recognition dictionary corresponding to the source of information being viewed (control target device) is selected. Next, the speech recognition processing unit 12b uses the selected dedicated recognition dictionary to compare and collate the utterance content (speech command) with each vocabulary (command) included in the selected dedicated recognition dictionary, The degree of match is calculated. Then, based on the calculation result, the “vocabulary” having the highest degree of match is determined as the command spoken by the user. Next, the device control signal generator 12c acquires the determined command and outputs a device control signal corresponding to the content of the command. The output device control signal is sent to the bus 7 by the CPU 12 as operation instruction data for the control target device.

  Hereinafter, an example of processing related to the identification of the speaker and the switching selection of the speech recognition dictionary based on the identification performed by the CPU 12 in cooperation with the DSP 11 in the in-vehicle speech recognition apparatus AR (FIG. 2) according to the present embodiment will be shown. This will be described with reference to FIG.

  First, as an initial state, each seat (driver's seat, front passenger seat, rear seat) user has already viewed information on a desired source (control target device), and the CPU 12 in the speech recognition unit 10 uses the management table (user It is assumed that information indicating the relationship between the seat where the user is seated and the source of the information viewed on the seat is created and stored in the memory unit 13.

  In the first step S1 in this state, the CPU 12 determines whether an utterance has been detected from the microphone array 9 via the DSP 11 (voice input unit 11a) (YES) or not (NO). If the determination result is YES, the process proceeds to the next step S2, and if the determination result is NO, the determination process is repeated until an utterance is detected. The “speech operation” described in parentheses in the processing content of step S1 will be described later.

  In the next step S2, the direction (seat) in which the user (speaker) who made the utterance is present is specified in the DSP 11 by the beamforming method using the microphone array 9 based on the control from the CPU 12. That is, the speaker is specified.

  In the next step S3, the recognition dictionary selection unit 12a in the CPU 12 refers to the management table stored in the memory unit 13, and from among a plurality of dedicated recognition dictionaries D1 to D3 stored in the HDD 8, A dedicated speech recognition dictionary corresponding to the source of information viewed by the speaker (for example, the navigation unit 5 for the driver's seat, the radio receiver 1 for the passenger seat, and the DVD player 3 for the rear seat). select.

  In the next step S4, the voice recognition processing unit 12b in the CPU 12 executes voice recognition for the utterance content (voice command) of the speaker using the selected dedicated recognition dictionary.

  In the final step S5, the device control signal generator 12c in the CPU 12 outputs a device control signal corresponding to the recognized command (utterance content), and executes the control corresponding to this on the control target device. . At that time, based on the control from the CPU 12, the display units 25 and 31 displaying the video related to the operation state of the control target device are controlled according to the utterance content (change of the screen, etc.) Control (sound change etc.) according to the content of the utterance is performed on the speaker 27 (including the wireless headphones 32) that outputs the sound related to the operation state of the control target device.

  As described above, according to the on-vehicle speech recognition apparatus AR according to the first embodiment, the user (direction in which the user is uttered) in the vehicle compartment is specified by the beam forming method (DSP 11) using the microphone array 9. Then, the CPU 12 creates a dedicated speech recognition dictionary corresponding to the source (control target device) of the information that the specified speaker is viewing from among the plurality of dedicated recognition dictionaries D1 to D3 stored in the HDD 8. I am trying to select it. In other words, a dedicated recognition dictionary in which vocabulary frequently issued by the speaker as a voice command to the control target device is registered is selected.

  Thus, when recognizing the utterance content (voice command) of the speaker, only the vocabulary registered in the selected dedicated recognition dictionary needs to be recognized, so that it is registered in the speech recognition dictionary as in the past. Compared with the case where speech recognition is performed for all vocabularies, the rate of misrecognizing vocabulary that does not match can be reduced. That is, the recognition rate for the voice command can be increased by performing voice recognition suitable for the speaker.

  For example, when the speaker is seated in the rear seat, the rate of misrecognition can be reduced by using the speech recognition dictionary D2 for only the DVD operation reproduced on the screen of the rear seat display unit 31. it can. In this case, even if the navigation information is displayed on the screen on the driver's seat side of the front seat display unit (dual display) 25, the navigation operation is not affected by the speech operation at the rear seat. The same applies when the speaker is seated in the passenger seat.

<Second Embodiment (see FIG. 5)>
In the on-vehicle speech recognition device AR (FIG. 2) according to the first embodiment described above, the direction of the sound source (the direction in which the speaker is present) by the beamforming method (DSP 11) using the microphone array 9 as means for specifying the speaker. ) Has been described as an example, but it goes without saying that the means for specifying the speaker is not limited to this. For example, when inputting an operation instruction by voice (speech), a certain switch or the like may be operated (speech operation), and this utterance operation may be detected by the CPU to identify the speaker. FIG. 5 shows the configuration of the in-vehicle speech recognition apparatus according to the embodiment in that case.

  The in-vehicle voice recognition device AR1 (FIG. 5) according to the second embodiment is different from the in-vehicle voice recognition device AR (FIG. 2) according to the first embodiment. The point where the utterance switch 50 is provided in the operation unit of each of the operation units 20 and 30, the point where the microphone 9a is provided instead of the microphone array 9, the point where the DSP 11 is omitted, and the CPU 14 is provided instead of the CPU 12, and this CPU 14 is a voice. The difference is that an input unit 14a, a speaker identification unit 14b, a recognition dictionary selection unit 14c, a speech recognition processing unit 14d, and a device control signal generation unit 14e are provided. Other configurations and functions thereof are the same as those in the first embodiment, and thus description thereof is omitted.

  In addition, the processing related to the speaker identification and the voice recognition dictionary switching selection based on the speaker identification performed by the CPU 14 in the second embodiment is basically the same as the processing according to the first embodiment (FIG. 4). Since there is, the description is omitted.

  In the in-vehicle voice recognition device AR1 according to the second embodiment, the same effects as those obtained in the in-vehicle voice recognition device AR according to the first embodiment described above can be obtained. Furthermore, in this embodiment, since the speaker can be easily identified by detecting the operation of the speech switch 50, the configuration is simpler than the case where the speaker is identified using the microphone array 9 and the DSP 11. And cost reduction.

<Third Embodiment (see FIG. 6)>
In the on-vehicle speech recognition apparatuses AR and AR1 (FIGS. 2 and 5) according to the first and second embodiments described above, a plurality of dedicated recognition dictionaries D1 to D3 are prepared in the HDD 8, and the function or speech switch of the DSP 11 is prepared. The case where one of the dedicated recognition dictionaries is selected corresponding to the information source (device to be controlled) being viewed by the speaker of the seat specified based on the operation of 50 has been described as an example. The form to change is not necessarily limited to this.

  Rather than selecting from a plurality of dedicated recognition dictionaries as described above, for example, when performing speech recognition on utterance content, depending on the source of information (control target device) that the specified speaker is viewing It is also possible to add “weighting” to prioritize the word to be recognized and recognize the recognized word to which the “weighting” is added as a command of the speaker. FIG. 6 shows an example of the speech recognition method in that case.

The in-vehicle speech recognition apparatus according to the present embodiment is not specifically illustrated, but basically has the same configuration as the in-vehicle speech recognition apparatuses AR and AR1 (FIGS. 2 and 5) according to the first and second embodiments. have. The difference in configuration is that the CPUs 12 and 14 do not have functional blocks corresponding to the recognition dictionary selection units 12a and 14c (however, the management table stored in the memory unit 13 is used) . Instead of preparing dedicated recognition dictionaries D1 to D3, one speech recognition dictionary shared by each source (control target device) is prepared, and a vocabulary (word) to be recognized for each source is set in advance. A table (WT1, WT2 in FIG. 6) that defines the relationship with weighting is prepared.

In the third embodiment, when performing speech recognition on the utterance content of the utterer specified by the CPU 12 (14), the utterer views the utterance by referring to the management table stored in the memory unit 13. The information source (control target device) is identified, and “weighting” is added only to the vocabulary corresponding to the control target device with reference to the tables WT1 and WT2. For example, when a navigation-related word “company” is spoken from the driver side and the passenger side via the microphone 9 (9a), the CPU 12 (14) recognizes the recognition word “ By adding weight (+10) only to “company”, “company” spoken from the driver's seat side is recognized as a voice command, and control corresponding to the recognized command is executed on the navigation unit 5. When the audio-related word “stop” is uttered from the driver side and the passenger side via the microphone 9 (9a), only the recognition word “stop” on the passenger side is weighted (+10). Thus, “stop” uttered from the passenger seat side is recognized as a voice command, and control corresponding to the recognized command is executed on the audio device (radio receiver 1, DVD player 3, etc.).

  In each of the above-described embodiments, the case where the in-vehicle voice recognition device AR (AR1) is incorporated as a part of the in-vehicle A / V / navigation system 40 has been described as an example, but the gist of the present invention (identifying the speaker, Change the recognition dictionary (select a dedicated speech recognition dictionary or change the weight of the recognition word) according to the information source (device to be controlled) that the speaker is watching and use the changed dictionary The voice recognition is performed, and control corresponding to the recognized utterance content is performed on the control target device), as is apparent from the above description, it is not necessarily incorporated into a system including both the A / V device and the navigation device. Of course, there is no need to use it.

  Further, in each of the above-described embodiments, the “management table” indicating the relationship between the seat where the user is seated in the passenger compartment and the source of information (control target device) viewed in the seat is stored in the voice recognition unit 10. Although the case where the data is stored in the memory unit 13 has been described as an example, it is needless to say that it is not always necessary to have it stored in the voice recognition unit 10 as is apparent from the gist of the present invention. For example, the management table is stored in the memory unit 24 in the H / U 20, and the CPU 12 (14) in the speech recognition unit 10 cooperates with the control unit 21 in the H / U 20 when necessary to The unit 24 (management table) may be referred to.

  Further, in each of the above-described embodiments, the case where the wireless headphone 32 is provided for the rear seat has been described as an example. However, the present invention is not limited to such a “wireless” type, and this headphone is also used when a headphone with a jack is used. The invention is equally applicable. In this case, the headphones are connected to the corresponding display unit 31 via a jack.

  In each of the above-described embodiments, the HDD 8 is used as a storage medium for storing the map data and the voice recognition dictionary. Instead, a DVD drive (DVD-ROM), a CD drive (CD-ROM), or the like is used. Other storage media may be used.

1 is a block diagram showing a configuration of an in-vehicle audio / video (A / V) navigation system incorporating an in-vehicle voice recognition device according to an embodiment of the present invention. It is a block diagram which shows a part of structure of the vehicle-mounted speech recognition apparatus which concerns on 1st Embodiment. It is a figure for demonstrating the method to identify the direction of a sound source (direction where a speaker exists) by the beam forming method using a microphone array in the vehicle-mounted speech recognition apparatus of FIG. It is a flowchart which shows an example of the process which concerns on the voice recognition dictionary switching selection etc. based on the identification of the speaker performed in the vehicle-mounted speech recognition apparatus of FIG. It is a block diagram which shows typically a structure of the vehicle-mounted speech recognition apparatus which concerns on 2nd Embodiment. It is a figure for demonstrating the method of the speech recognition performed in the vehicle-mounted speech recognition apparatus which concerns on 3rd Embodiment.

Explanation of symbols

1-6 ... Source of information (control target device) that the speaker is watching,
8 HDD (dictionary storage means),
9: Microphone array (voice input means),
9a: Microphone (voice input means),
10 ... Voice recognition unit,
11 ... DSP (speaker identification means),
12, 14 ... CPU (control means),
13: Memory unit (table storage means),
20, 30 ... operation unit,
25, 31 ... display unit (display means),
27 ... Speaker (voice output means),
32. Headphone (sound output means),
40. Car audio / video (A / V) navigation system,
50. Utterance switch (speaker identification means),
AR, AR1 ... Vehicle speech recognition device,
D1, D2, D3 ... speech recognition dictionary (corresponding to each control target device),
P1, P2, P3 ... passengers (users) in the passenger compartment
WT1, WT2 ... A table defining the relationship between recognized words and weights.

Claims (10)

Voice input means for voice input of information instructed by the user in the passenger compartment;
A speaker identifying means for identifying a user who has spoken via the voice input means;
Stores one speech recognition dictionary in which vocabulary related to each operation instruction of a plurality of control target devices is registered, and the relationship between the vocabulary to be recognized for each control target device and a preset weighting A dictionary storing means for storing the defined first table;
Memory means storing a second table that defines a relationship between a seat on which a user is seated and a control target device that is a source of information viewed by the user of the seat;
Control means operably connected to the voice input means, speaker identification means , dictionary storage means and memory means ;
When the control means specifies a speaker in cooperation with the speaker specifying means, the control means refers to the voice recognition dictionary and the first and second tables, and the speaker is watching Add a predetermined weight to the vocabulary corresponding to the controlled device that is the source of information, perform speech recognition on the utterance content of the speaker by referring to the vocabulary with the weight added, and according to the recognized utterance content A vehicle-mounted speech recognition apparatus, wherein the control target device is controlled. The voice input means is a microphone array installed at a predetermined location in the passenger compartment.
The speaker specifying unit receives a signal having a high sound pressure level based on the direction signal based on the signal detected by each microphone of the microphone array and generating a beam signal. The vehicle-mounted speech recognition apparatus according to claim 1, further comprising: a unit that identifies a direction in which the speaker is speaking as a direction in which the speaker is present. The voice input means is a microphone installed at a predetermined location in the vehicle interior,
The in-vehicle speech recognition apparatus according to claim 1, wherein the speaker specifying means is an utterance switch operated when a user utters via the microphone. A plurality of display means provided corresponding to each seat in the passenger compartment,
When the control unit performs control according to the recognized utterance content on the control target device, the control unit applies the utterance content to the display unit displaying an image related to the operation state of the control target device. The vehicle-mounted speech recognition apparatus according to claim 1, wherein control is performed in accordance with the control. A plurality of audio output means operably connected to the plurality of display means,
When the control unit performs control according to the recognized utterance content on the control target device, the control unit outputs the utterance content to the voice output unit that outputs the voice related to the operation state of the control target device. The vehicle-mounted speech recognition apparatus according to claim 4 , wherein control is performed according to In a vehicle speech recognition device having a function of performing speech recognition corresponding to an operation instruction for a control target device spoken by a user in a vehicle interior,
A first that defines a relationship between a vocabulary to be recognized for each control target device and a preset weight, together with one speech recognition dictionary in which vocabulary related to each operation instruction of a plurality of control target devices is registered . And a second table that defines a relationship between a seat on which the user is seated and a control target device that is a source of information viewed by the user of the seat, in the storage unit,
Identify the speaker when the utterance is detected,
Referring to the voice recognition dictionary and the first and second tables, a predetermined weight is added to the vocabulary corresponding to the control target device that is the source of the information that the speaker is viewing,
Performing speech recognition on the utterance content of the speaker by referring to the weighted vocabulary;
A speech recognition method, wherein control according to the recognized utterance content is performed on the control target device.   The speech recognition method according to claim 6, wherein a speech is detected by a microphone array, and when the speech is detected, a direction in which the speaker is present is specified by a beam forming method using the microphone array.   The speech recognition method according to claim 6, wherein an utterance is detected by a microphone and the speaker is specified based on an operation of an utterance switch.   When the control according to the recognized utterance content is performed on the control target device, the control according to the utterance content is performed on the display unit displaying the video related to the operation state of the control target device. The speech recognition method according to claim 6.   When the control according to the recognized utterance content is performed on the control target device, the control according to the utterance content is performed on the voice output unit that outputs the voice related to the operation state of the control target device. The speech recognition method according to claim 6, wherein the speech recognition method is performed.

Images