JP2002258892A - Operation unit for voice-recognizing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice-recognizing device operation unit that can perform a desired operation by estimating a directive to a function desired by a user only by speaking directive, in the case a particular directive that operates device functions are commonly used for a plurality of functions, and when all of these devices operate. SOLUTION: When a voice operation instruction merely called 'up' is inputted from a microphone 2 and recognized by a voice-recognizing part 4, a function word/ directive-identifying part 5 analyzes the voice operation instruction to identify that the voice operation instruction is composed only of a directive. In such a case, an operable function detecting part 6 reads devices that are currently operating from an operating device storing part 21 and retrieves functions which are made to operate by the 'up' within the range of the devices from an operation instruction/use frequency database 12, and an operation function estimating part 7 reads use frequency data corresponding to the directive of each function in the database, when a plurality of functions exist, selects a function having the highest use frequency estimating that the function is an operation function desired by the user, forms an operation word, and subsequently converts the operation word into a device operation signal to be outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recognition device operation device which enables a user to recognize a device operation voice input from a microphone and to operate various devices. Speech recognition device operation device that enables a specific device to be operated using a common language for each device or function without specifying the device or function to be operated when the function is operating About.

[0002]

2. Description of the Related Art Recent advances in voice recognition technology have made it possible for a user to recognize voice input from a microphone and operate various devices. In particular, in the case of a vehicle-mounted device, since the person who operates the device is mainly a driver, use the voice recognition device operation device so as to distract the gaze by operating these devices and to ensure that safe driving is not impaired by inattentive driving. In many cases, a voice input from a microphone installed near a driver's mouth or around a driver's seat is recognized, and, for example, an operation of a navigation device, an audio device, a mobile phone, and an operation of an air conditioner or the like are performed. It is also being developed.

[0003] Also, in general households, a study has been conducted to operate various devices such as a television, a video, an audio device, an air conditioner, a lighting device, a hot water supply device, a door lock and the like using a voice recognition device operating device by voice. It is being advanced by equipment manufacturers. In this way, by performing device operations by voice recognition in ordinary households,
Many remote controls for each TV, video and audio device,
Or remove the remote control of the air conditioner so that the lighting can be turned on and off and the brightness can be adjusted without going to the place where the switches are installed. In addition, the use of voice recognition equipment operating devices has been studied in many fields, such as enabling door locks to be performed from a distance and limiting the unlocking of door locks to only specific persons by registering specific voices. In particular, with the increase in the number of elderly households, it is about to become widely used as a device for handicapped people.

[0004] On the other hand, for example, with regard to audio equipment for vehicles, a medium such as a radio, a cassette tape, a CD, an MD, and a CD changer or an MD changer are connected to a head unit, and a medium selected from these is selected. Is output to the head unit. The head unit converts this audio signal to DS
By performing audio signal processing such as P, an equalizer function and a sound field control function are performed, or the volume is adjusted by an amplifier function and output from a speaker.

When such a speech recognition device operating device is used for such an audio device, for example, in order to adjust the volume of the audio device, for example, a word such as "volume up" is uttered. The speech recognizer then processes this word by speech, then converts it into text data and analyzes its contents, and interprets the above word as a word indicating the function of "volume" and how to change that function. Break it down into the word "up" to instruct.

[0006] The word indicating the function as described above is hereinafter referred to as a "function word", and the word indicating how to change the function is referred to as a "instructive word".

[0007] Accordingly, the word "volume" is recognized, the operation of the volume adjustment function is selected from various operation functions of the audio device, and the word "up" is recognized, and the operation of increasing and decreasing the volume in the function is recognized. Among the operations to be performed, an operation in the direction of increasing the volume is selected and performed, and the operation of increasing the volume of the audio device is performed by recognizing the above words by voice.

Also, when the operation instruction of the audio device is given by using the voice recognition device operation device, when the receiving station is changed when the radio is selected from various audio devices, the receiving frequency is changed in a direction of increasing the receiving frequency. At this time, for example, an operation instruction is given by uttering a word such as “frequency up”. The speech recognition device operation device performs speech recognition processing on this word, and decomposes the word into a function word “frequency” and a descriptive word “up”. Where "frequency"
By selecting the operation of the function that changes the tuning frequency among the various functions of the radio receiver by recognizing the word, the operation of increasing and decreasing the tuning frequency in that function by recognizing the word “up” Among them, the operation in the direction of increasing the frequency is selected and the operation is performed, and thus the above-mentioned words are recognized by voice, thereby selecting the radio receiving station in the direction of increasing the frequency.

[0009]

As described above, when operating the audio equipment, particularly, the increase and decrease of the volume and the increase of the reception frequency are performed, for example, by selecting media such as radio, CD, MD, and the like for AM broadcast and FM broadcast. Unlike selecting and performing an operation like selecting one, for example, for volume, specify the increase with the word "up", and then specify the increase with the word "up" according to the result. In order to continuously perform this operation, there is a device that continuously instructs operations for the same function. In this regard, the same applies to the radio reception frequency as described above, and when increasing the reception frequency from the currently receiving station,
When the station selected by the utterance of “up” as described above is not the desired station, the utterance of “up” is continuously performed again, and this is repeated or the utterance of “down” is performed. Become.

[0010] In the case where the operation instruction is continuously given as described above, when the operation of increasing the volume is desired to be performed in the above-described case, the word "up" is used for the function of increasing the volume. And the function of increasing the reception frequency, the words "volume up" and "volume up" must be repeated to clearly indicate which function is to be improved. This is the same when performing the operation of increasing the frequency. In order to increase the frequency, it is necessary to repeat the words “frequency up” and “frequency up”.

In addition, such a phenomenon is not limited to the above-described audio equipment. For example, when adjusting the temperature setting of an air conditioner, the "temperature is increased" in order to increase the set temperature.
The same is true when saying the word, and when a voice recognition device operating device is used as a general household device, if the lighting device can be operated by voice in addition to the audio device and the air conditioner as described above. It is necessary to repeat words such as “increase in light amount” and “increase in light amount” in order to adjust the light amount.

As described above, when there are a plurality of functions for performing continuous operations using the same descriptive word, simply uttering only the descriptive word cannot be distinguished from the operation of other functions, so that it is always uttered together with the functional word. There is a need to. However, for example, the word “up” is frequently used in vehicle-mounted devices, for example, when the radio is operating in an audio device, and generally the volume of the audio even when the air conditioner is operating is often used. Only when the other equipment is not working, is there only a loudness. Also, 1
When the word "up" is spoken continuously after the word "up" is used for the volume only once, it is very likely that the volume is loud. Nevertheless, in the conventional voice recognition device operation device, the function word and the instruction word must always be uttered together, which is inconvenient.

[0013] As a countermeasure, for example, when performing an operation to increase the volume, the instruction word “up” is used, and when performing the operation to increase the frequency, the instruction word “up” is used. It is preset to use different directives for each function, such as using the directive "cold" for the operation of increasing the amount of light and using the directive "bright" for the operation of further increasing the amount of illumination. It is also conceivable to keep it.

However, for each function, only the operation of increasing the amount of setting of the function is performed, and it is apparent to the user that the same operation is performed using different words for each function. It may be difficult to operate, and it may be a cause of confusion in operation.

Therefore, the present invention is applied to a case where a specific instruction for operating a function of a device is commonly used for a plurality of functions, and even when all of these devices are operating. A main object of the present invention is to provide a speech recognition device operation device that can estimate a command word for a function desired by a user just by uttering the command word and perform a desired operation.

[0016]

In order to solve the above-mentioned problems, a voice recognition device operating device according to the present invention has a voice recognition unit for recognizing a device operation voice input by a user, and a voice recognition unit that is recognized by the voice recognition unit. An identification unit that identifies a function word that selects a function of the device and a command word that instructs the operation of the function from the voice, and detects that a voice input of the command word alone is performed;
An operation instruction database that records a function of a device and an instruction to operate the same, and an operation of detecting whether the instruction identified by the identification unit is an instruction common to a plurality of functions from the operation instruction database. An enabling function detection unit, a usage frequency database that records usage frequency corresponding to the instruction word, and the usage frequency when the operable function detection unit detects that the instruction word is common to a plurality of functions. A use frequency search unit for searching the database for the most frequently used function of the plurality of functions, and adding a function word of the most frequently used function searched for by the use frequency search unit to the input descriptive word; An operation word forming unit that outputs the operation word,
A usage frequency correction unit configured to correct data in the usage frequency database when an operation instruction signal for the device is output.

Further, another voice recognition device operation device according to the present invention includes an active device detection unit that detects an operation state of a connected device connected to the voice recognition device operation device, and the input instruction word indicates the activation. When the instruction word common to a plurality of functions in the active device detected by the middle device detection unit is used, the use frequency search unit searches for the use frequency of the plurality of functions.

Further, the other voice recognition device operation device according to the present invention detects the output operation instruction signal, and subsequently inputs an instruction word capable of performing the same operation as the operation of the previously output function. In such a case, there is provided means for selecting the same function as the previous output, prior to other functions.

Further, another operation apparatus for a voice recognition device according to the present invention is arranged such that, when the same operation word is continuously output, the operation amount of the device operated by the operation word is changed according to the degree of continuation. It was made.

Further, in another voice recognition device operating device according to the present invention, the use frequency database records values obtained in advance by use of model users as initial values.

Further, another voice recognition device operating device according to the present invention uses the voice recognition device operating device for operating a vehicle mounted device.

[0022]

Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a functional block diagram showing main functional blocks and their mutual relationship in an embodiment of a voice recognition device operating device according to the present invention. In FIG. 1, when a voice for instructing a user to operate a device is input from the microphone 2 to the voice recognition unit 4 of the voice recognition device operation device 1 after the voice input start switch 3 is operated, the voice recognition unit 4 outputs the voice data. Is analyzed, the data in the speech recognition dictionary database 22 is searched for the analyzed data, and the voice of the user is converted into text data.

The recognized voice is a function word / indicative word identification unit 5
In terms of words uttered by the user, such as "volume up", the presence of the function word "volume" and the instruction word "up" are identified, and the voice of the operation instruction input by the user is It is determined whether or not the voice is combined. As a result, when it is determined that both the function word and the descriptive word are included as described above, the word is a word that can be used to operate the device as it is, and the subsequent processing is performed. And outputs the operation instruction to the operation instruction signal output unit 9. In the illustrated embodiment, the operation instruction signal output unit 9 shows an example in which three devices, ie, operation devices A to C, which operate by voice recognition are connected. On the other hand, an operation instruction to increase the volume by one level is issued. It should be noted that the amount of increase in this one stage can be set so as to change according to the output status of the voice operation instruction as described later.

When the function word / instructive word discriminating unit 5 identifies that a speech which is only an instructive word such as "up" and has no function word is input, it is sent to the operable function detecting unit 6. Output. Among the operating devices A to C that can be operated by voice recognition, those that are actually in operation and are in a state where they can perform their functions are detected by the operating device detection unit 20, The detection signal is stored in the operating device storage unit 21. In the operable device detection unit 6, as described above, "up"
When only such an instruction word is input, the data in the operating device storage unit 21 is searched to specify the devices and functions that can be currently actually operated by voice recognition.

As shown in FIG. 2, for example, as shown in FIG. 2, the voice recognition device operation device 1 has an operation instruction usage frequency that stores the frequency of use of the operation device, its functions, and various operations on the functions. A database 12 is provided. This figure shows an example of a device mounted on a vehicle that can be operated by voice, and shows an example in which a head unit, a radio, a CD player, etc. are provided in an audio device as examples of those devices. An air conditioner, a navigation device, and the like are provided as on-vehicle devices that can be operated by voice.
The example which can be operated by voice by one voice recognition device operation device is shown. The operable function detection unit 6 further searches the operation instruction / use frequency database 12 for the currently operating device detected by the operating device storage unit 21 and operates the device with the same instruction. Find out if the feature exists elsewhere, and if so, what it is.

In the operation instruction / use frequency database 12 shown here, data indicating what instruction words can be used for operations in various functions of each device as described above is recorded, and thereby specified. It is possible to search for what kind of equipment can be operated by the indication of the above, and to find out what equipment or function is operated by the same indication, and in addition to this indication Record usage data for each word,
This shows an example in which data for comparing the frequency of use of each descriptive word with others is recorded, and various searches can be performed. However, this is divided, for example, the data of the use frequency is recorded in the memory as another use frequency data together with the corresponding instruction word, and the data other than the use frequency data is recorded in the other memory as the operation instruction data, and each is independently executed. It can be configured to search and read.

In the head unit of the audio equipment in FIG. 2, there are a volume, an equalizer, etc. as an example of a function which can operate the head unit. Among the functions of the volume, an instruction which can be operated by voice is a volume. This instruction is "up"
-The example which can be performed by "down" is shown. In addition, for the equalizer, operation instructions for operating the equalizer function provided in this audio device are issued according to the equalizer function, and the instruction word at that time is set to a user-friendly one in accordance with the operation instruction. You.

On the other hand, in the radio of the same audio equipment, an example of a reception frequency is shown as an example of a function that can be operated by voice, and the words “up” and “down” are used when an instruction to increase or decrease the reception frequency is given. It is performed by using. In the same figure, C
The D player also shows an example in which the playback function is operated by voice. The playback operation is instructed by the word "play" when instructing the playback, and the stop is called "stop", and when skipping to the next song, it is called "skip". Words indicate that each operation can be performed. In addition,
Even when an MD player is provided as an audio device, similar functions can be achieved by using similar words.

FIG. 2 shows an example in which the air conditioner can also be operated by voice. In this air conditioner, the functions of changing the set temperature and changing the air volume are operated. "Up" instructions
An example in which the operation is performed by the word “down” is shown. In addition, many functions of the navigation device are operated by various voice instructions as already widely performed. In addition to the above example, various devices can similarly issue operation instructions by using predetermined sounds.

In the example shown in FIG. 2, among many voice operation devices that can be operated by voice, the devices and operations that can increase or decrease their functions by the words “up” and “down” In the head unit of the audio equipment, the volume is increased or decreased, the reception frequency of the radio is increased or decreased, the set temperature of the air conditioner is increased or decreased, and the air volume of the air conditioner is increased or decreased. An example in which the user cannot perform an operation by simply inputting the voice of “up” without knowing which device or function the user is going to perform the operation instruction.

The operation instruction / use frequency data table of FIG. 2 includes a use frequency recording unit corresponding to each of the above-mentioned instruction words. The use frequency data is stored in the operation instruction / use frequency database 12 of FIG. As shown, the use frequency initial data 1 according to the use results by the model user in advance
0 has been entered. Every time the user actually gives an operation instruction by voice, the operation signal output from the operation instruction signal output unit 9 is detected by the operation signal detection unit 15, and the data is rewritten by the use frequency correction unit 16, whereby the data is rewritten. An example in which the usage frequency data 11 is formed is shown.

Based on this data, for example, when the user omits the utterance of the function word such as “volume” and simply utters “up”, the use frequency search unit 13 uses the operation instruction / use frequency database 12 The frequency data 11 is searched, and the most frequently used function is determined to be the function intended by the user, and is output to the operation function estimating unit 7. In this determination, an appropriate data is selected based on not only data obtained by using a preset model user but also data that also takes into account the personality of the user.

Based on the data of the function that can be operated by the user's voice obtained by the operable function detection unit 6 as described above, the operation function estimation unit 7 uses only the operating devices for the operating frequency. The search unit 13 performs the above-described search, and searches for devices and functions that are estimated to be most intended by the user. Thus, for example, in the example shown in FIG. 2, when the audio device is activated and the air conditioner is not activated, when searching for a function corresponding to the instruction word “up”, the data of the air conditioner is not searched, and only the audio device is searched. Perform a search for

Further, in the case where the selected audio device connected to the head unit which always operates in the audio device is, for example, a radio, the same as the instruction word of the word for increasing the frequency among the words for operating the function of the radio. A search is made to find that "up" is being used, and a use frequency n21 of "up" for operating in a direction to increase the radio reception frequency and a use frequency n11 of "up" for increasing the volume of the head unit are obtained. In comparison, if the volume “up” is used more frequently, it is estimated that the user is going to operate the volume function. Based on the result, the operation word forming unit 8 in FIG. 1 adds the function word “volume” to the instruction word “up” to form an operation word “volume up”. As a result of estimating the function in this way and forming and outputting the operation word, if the user's intention is different from the user's intention, the user notices this and, as in the past, for example, "frequency up" By uttering an operation word to which a function word for performing an intended function is also added, it is possible to execute the conventional operation.

In the above case, the radio is not selected and C
When the operation of the D player is selected, this is obtained from the data in the operating device storage unit 21. Therefore, the use frequency search unit 13 searches the operation instruction use frequency database 12 under this condition. As a result, it is found that there is no device or function that can be operated “up” among the currently operating devices other than the volume of the head unit, and an operation instruction to increase the volume is immediately issued.

On the other hand, when the radio is selected and operating in the audio equipment, and when the air conditioner is also operating,
When the word "up" is issued, the data of the air conditioner is searched in addition to the audio equipment as described above, and the word "up" is also used for the set temperature and the air volume. For all of these function operations, a search is made for the use frequency at which "up" is used, and the most frequently used one is selected. At this time, n11 and n2 of the use frequency data shown in FIG.
The data of 1, n31, and n41 are compared, and the largest one is selected. As a result, when it is determined that the volume is the most frequently used, a device operation signal corresponding to the word “volume up” is output to the audio device.

In the example shown in FIG. 1, a latest output storage unit 14 is further provided to store the latest operation words output from the operation instruction signal output unit 9, and the operation function estimating unit 7 performs the above-described functions. In addition to performing the word estimation operation, referring to the data in the latest output storage unit 14, when the previously output one is, for example, "up" of the radio frequency, the user utters the next successive sound. "Up" shows an example in which it is naturally determined that it is the same as the previous one, and an operation of increasing the frequency can be performed without searching for the use frequency data.

The use frequency data 11 in the operation instruction / use frequency database 12 is obtained by actual use by the model user as described above. In this case, for example, the use frequency database shown in FIG. This can be done by the device 30. That is, in the illustrated example, a device substantially similar to a conventionally used voice recognition device operating device is used, and a voice recognition unit 33 detects a user input from the microphone 31 after operating the voice input start switch 32. The voice is recognized with reference to the voice dictionary database 34 and output to the operation word identification unit 35.

The operating state of each of the operating devices A to C is detected by the operating device detecting section 44 and is temporarily stored in the operating device storage section 45. The operation word identification unit 35 identifies the operation word with reference to the data in the operating device storage unit 45,
This is used as an apparatus operation signal by the operation signal output unit 36 and output to a predetermined operation apparatus to operate the operation apparatuses A to C. At this time, the voice input by the user is executed by uttering an accurate operation word in which the function word and the instruction word are integrated as in the related art.

However, when data is accumulated by use of the model user, the same system as that shown in FIG. 1 is used, and only the operation instruction / use frequency database initially has no use frequency data, or an appropriate The model user should use the data with the values entered, and be aware of the fact that the model is being used for data storage in this system, and use various trial and error to store the data. Is also possible.

It should be noted that the operating devices connected here are usually those connected to various devices that are standard equipment of a specific vehicle, and a large amount of data may be obtained in various usage modes. In addition, for example, when a car audio equipment maker provides a function according to the present invention to audio, in particular, only the audio equipment can be operated by voice, and the volume “up” and the frequency “ The usage frequency data of "up" can be obtained and used, or can be used in combination with another specific device to obtain data.

The above-mentioned operation word output from the operation signal output unit 36 is detected by the operation signal detection unit 39, and the operation instruction / use frequency database 38 having the same data structure as that shown in FIG. Frequency data 37
Is updated by the use frequency updating unit 40. At the time of this update, the data can be updated by adding one to the usage frequency data every time the operation instruction by voice is performed once. For example, when the operation instruction by voice is performed once, for example, a predetermined count is used. It is also possible to adjust the update speed by adding a value multiplied by 0.2 or the like. Further, the value can be increased along an exponential curve in order to prevent the value from becoming extremely large. It should be noted that such updating of the database may be similarly performed in the use frequency correction unit 16 in the apparatus shown in FIG.

In the above-mentioned use frequency database creating apparatus composed of a block diagram as shown in FIG. 4, processing can be performed according to, for example, an operation flow shown in FIG. That is, in the process of creating the usage frequency database, input of operation word voice is received (step S21). This is performed by inputting a signal for starting use of the voice recognition device from the voice input switch 32 to the voice recognition unit 33 in FIG. Next, among the operating devices A to C connected so that the operation can be performed by the voice recognition, the currently operating device is detected by the operating device detecting unit 44, and the data is operated. It is stored in the middle device storage unit 45 (step S22).

Next, referring to the data in the operating device storage section 45, a device for performing an operation instruction is searched for and specified (step S23), and an input operation word is specified (step S2).
4) The operation signal output unit 36 outputs a device operation signal to a predetermined device. The signal output at this time is detected by the operation signal detecting section 39, and the corresponding use frequency data 37 in the operation instruction / use frequency database 38 is updated by the use frequency update section 40 (step S25), and this processing is terminated (step S25). Step S26). In the updating, in addition to simply adding 1 to the data, an updating process can be performed by various updating methods such as adding a value obtained by counting as described above or increasing exponentially.

As a result, in the data table shown in FIG. 4B which exemplifies only the devices or functions that operate with the above-mentioned instruction word "up", for example, the operation word "volume up" is output. Then, the use frequency data n11 corresponding to the operation instruction to increase the volume in the head unit of the audio device is updated. Similarly, for example, when the operation word “temperature up” is output, the use frequency data n31 corresponding to the operation instruction to increase the set temperature of the air conditioner is updated. After this update is performed by the model user for a predetermined period, this data is written into the memory when the product is manufactured, and is used as an initial value when the user first uses the product.

As described above, when a plurality of devices or functions are performed by the same descriptive word, data for estimating which device or function the user intends is obtained in advance by using the model user. Since the initial value that seems appropriate is entered, the user of this product will operate almost appropriate devices or functions even when the user inputs only the instruction word such as “up” as described above from the beginning. be able to.

In the device shown in FIG.
When a product having the initial data obtained as described above in the usage frequency database 12 is used, for example, processing is sequentially performed according to an operation flow as shown in FIG. An appropriate function word can be added to the operation word only by uttering the instruction word, and the operation word can be output. That is, in the process of determining the instruction word with a function word, the input of the voice of the operation word for operating the various devices by voice is first received (step S1). This is performed by inputting a signal for starting use of the voice recognition device from the voice input switch 3 to the voice recognition unit 4 in FIG.

Next, it is determined whether or not the input operation word is a single instruction word (step S2). The discrimination is performed by analyzing the operation instruction signal converted into text data by the voice recognition unit 4 in the function word / instruction word identification unit 5 in FIG. When it is determined in step S2 that the operation word is the only instruction word, that is, when it is determined that the operation word is uttered only by the instruction word as in the above-described "up", detection of the operating device in the next step S3 is performed. The storage is performed. This is performed in the active device detection unit 20 and the active device storage unit 21 in FIG. On the other hand, if it is determined that the operation word is not a single instruction word, the operation word immediately proceeds to step S10 because the operation word is a completed operation word in which the function word and the operation word are combined, such as “voice up”. Output as an instruction signal.

Next, the operation target device is specified based on the data in the operating device storage unit (step S4). This is performed by the operable function detection unit 6 in FIG. 1, and for example, it is detected that the air conditioner, which is a device that can be operated by voice, is not currently used. After that, a function operable with the instruction word is obtained from the operable function database (step S5). In the embodiment shown in FIG. 1, the operable function database corresponds to the operation instruction / use frequency database 12, and the operable function detection unit 6 searches for this data.

Thereafter, it is determined whether there is only one operable function (step S6). In this judgment, 1
If it is determined that the type is only the type, the function word is added to the instruction word to make it an operation word (step S13), and this is output as an operation instruction signal (step S10). This is performed in the operation word forming unit 8 and the operation instruction signal output unit 9 in FIG. As described above, among the currently operating devices, when it is determined that there are a plurality of devices or functions that operate with the same instruction, in the illustrated embodiment, it is determined whether the instruction is the same as the previous one. Yes (Step S
7).

In this discrimination, when the same descriptive word as the previous one is continuously uttered, it is estimated that the user has continuously uttered the same descriptive word in order to perform the same function as the previous operation word. Then, the same function word as the previous one is automatically selected (step S12), this function word is added to the instruction word to make it an operation word (step S13), and this is output as an operation instruction signal (step S10). Such comparison with the previous instruction word can be performed by searching data stored in the latest output storage unit 14 in FIG.

When it is determined in step S7 that the instruction word is not the same as the previous instruction word, a function that is frequently used is searched and determined as a function word (step S8).
This is performed by the use frequency search unit 13 in FIG. Thereafter, the instruction word is added to the function word of the determined function to make it an operation word (step S9), and this is output as an operation instruction signal (step S10).

After the operation instruction signal determined according to various cases as described above is output, the frequency of use of the instruction word of the determined function is updated (step S11). This updating can be performed by adding one to the usage frequency data, similarly to the updating of the usage frequency data by the use of the model user in FIGS. 4 and 5, and adding a value obtained by multiplying the other one by a coefficient. Further, by gradually changing this coefficient, the coefficient can be increased along an exponential curve, and an updating method such as preventing data overflow can be adopted. When such data updating is completed, the operation flow ends (step S14).

The present invention can be embodied in various forms other than the above embodiment. For example, when the user uses the voice recognition device operation device, for example, the operation of "up" the volume is continuously performed. When the repetition is found, for example, by detecting that the same output is continuously stored in the latest output storage unit 14 in FIG. 1, this data is separately recorded in a memory, and the volume is increased, for example. When the operation instruction is given, it is estimated that the same operation instruction has been repeatedly given because the amount of increase in volume in the audio device is smaller than what the user intended, so that the data in the memory will be By increasing the signal of the operation amount output from the operation instruction signal output unit 9 based on this, the volume of the sound may be increased in accordance with the intention of the user. At this time, the degree of increase in the volume can be changed according to the degree to which the user repeats the same instruction word. In addition, the operation instruction signal output unit can be operated by various means by performing appropriate learning, such as outputting a signal of increasing the volume twice, for example, by uttering “up” once according to the degree, As a result, it is possible to perform operation control according to the personality of the user.

In the above-described embodiment, an example in which the present invention is used particularly for operating equipment mounted on a vehicle is shown. In a similar manner, it can be widely used for device operation by voice recognition in various fields, such as voice operation of the device and comprehensive operation thereof.

[0056]

The present invention is configured as described above.
Even if a specific directive to operate the function of the device is used in common for multiple functions, by estimating the function desired by the user by use frequency data, just uttering the specific directive, A function word that instructs this function and the above-mentioned instruction word are combined into an operation word, and an appropriate device operation can be performed. As a result, in all cases, in principle, there is no need to utter a functional word, and an easy-to-use voice recognition device operation device can be provided.

Further, another voice recognition device operating device according to the present invention includes an active device detecting section for detecting an operating state of a connected device connected to the voice recognition device operating device, and the input instruction word is used for the operation command. When the instruction word common to a plurality of functions in the active device detected by the middle device detection unit is used, the use frequency search unit searches for the use frequency of the plurality of functions. Can be presumed to be that of the currently operating device, so that the function intended by the user can be more reliably estimated. In addition, since the search range of the use frequency can be limited to the currently operating device, even if there is a function that operates with many common descriptive words on the data, there is actually only one function. At least the search range is narrowed, so that high-speed and reliable search can be performed.

Further, the other voice recognition device operating device according to the present invention detects the output operation instruction signal, and subsequently inputs an instruction word capable of performing the same operation as the operation of the previously output function. In such a case, a means for selecting the same function as the previous output is provided prior to other functions, so that the operation of the function previously operated by voice recognition can be performed by simply speaking only the descriptive word. An operation can be performed, and an easy-to-use voice recognition device operation device can be provided.

Further, in another voice recognition device operating device according to the present invention, when the same operation word is continuously output, the operation amount of the device operated by the operation word is changed according to the degree of continuation. Therefore, for example, when it is detected that the volume is frequently increased continuously, the volume is increased greatly by one operation instruction, and the same operation instruction is given in accordance with the personality of the user. The device can be operated, and a speech recognition device operation device with good usability can be provided.

In another voice recognition device operating device according to the present invention, the use frequency database previously records a value obtained by use of a model user as an initial value. From the time of use,
Even when a descriptive word is simply input, an almost appropriate function can be selected and an operation instruction can be given.

Further, in the other voice recognition device operating device according to the present invention, since the voice recognition device operation device is used for operating the vehicle mounted device, the operation of the vehicle mounted device which is usually performed by the driver can be safely performed. When performing voice recognition for driving, an appropriate operation instruction can be given only by the instruction word without considering the combination of the function word and the instruction word, and the device can be easily used.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of an embodiment of the present invention.

FIG. 2 is a table showing an example of data recorded in an operation instruction / use frequency database used in the embodiment.

FIG. 3 is an operation flowchart of the embodiment of the present invention.

FIG. 4 shows usage frequency data recorded in an operation instruction / frequency database used in the embodiment of the present invention.
It is a block diagram showing an example of a use frequency database preparation device for obtaining the initial data.

FIG. 5 is an operation flowchart for executing the usage frequency database creation device.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 Voice recognition device operation device 2 Microphone 3 Voice input start switch 4 Voice recognition unit 5 Function word / instruction word identification unit 6 Operable function detection unit 7 Operation function estimation unit 8 Operation word formation unit 9 Operation instruction signal output unit 10 Frequency of use Initial data 11 Usage frequency data 12 Operation instruction and usage frequency database 13 Usage frequency search unit 14 Latest output storage unit 15 Operation signal detection unit 16 Usage frequency correction unit 17-19 Operating device 20 Active device detection unit 21 Active device storage unit 22 Speech recognition dictionary database

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06F 17/60 176 H04Q 9/00 331A G10L 15/00 G10L 3/00 537C 15/28 551Q H04Q 9/00 331 551G

Claims (6)

[Claims] A voice recognition unit for recognizing a device operation voice input by a user; a function word for selecting a function of the device from a voice recognized by the voice recognition unit; and a command word for instructing an operation of the function. And an identification unit that detects that the voice input of the instruction word alone has been performed; an operation instruction database that records the function of the device and the instruction word for operating the device; and the instruction word identified by the identification unit is An operable function detection unit that detects whether or not the instruction word is common to a plurality of functions from the operation instruction database; a usage frequency database that records a usage frequency corresponding to the instruction word; A use frequency search unit that searches the use frequency database for the most frequently used function from the use frequency database when the detection unit detects that the instruction word is a common descriptive word for the plurality of functions; An operation word forming unit that adds the function word of the most frequently used function searched by the degree search unit to the input instruction word and outputs the operation word as an operation word; A voice recognition device operating device, comprising: a use frequency correction unit that corrects data in a database. 2. The apparatus according to claim 1, further comprising: an operating device detecting unit that detects an operating state of the connected device connected to the voice recognition device operating device, wherein the input instruction word is an operating device detected by the operating device detecting unit. 2. The speech recognition device operating device according to claim 1, wherein when the instruction word is common to a plurality of functions, the use frequency search unit searches for the use frequency of the plurality of functions. 3. An output operation instruction signal is detected, and when an instruction word capable of performing the same operation as the operation of the previously output function is successively input, priority is given to other functions.
2. The apparatus according to claim 1, further comprising means for selecting the same function as the output. 4. The operation of a speech recognition device according to claim 1, wherein when the same operation word is continuously output, the operation amount of the device operated by the operation word changes according to the degree of the continuous operation word. apparatus. 5. The voice recognition device operating device according to claim 1, wherein the use frequency database records values obtained by using a model user in advance as initial values. 6. The voice recognition device operating device according to claim 1, wherein the voice recognition device operation device is used for operating a vehicle mounted device.