KR100677711B1 - Voice recognition apparatus, storage medium and navigation apparatus - Google Patents

Abstract

Provide a speaker with a voice recognition device that is as convenient to use as possible.

When the voice is input (S120), the voice is analyzed to select a plurality of candidate single syllables (S125), and the first candidate single syllable is notified (S135). Further, the processing is branched by the operation of the confirmation SW or by input of the voice (S140). When the confirmation SW is operated, the notified candidate single syllable is determined as the confirmation single syllable (S145). The candidate single syllable is selected by analyzing the voice (S125). As a result, in the case where the recognition cannot be made accurately, the speaker only needs to continuously re-ignite without operating the switch or the like and giving a re-input instruction.

Single syllable, speech recognition, GPS, navigation

Description

VOICE RECOGNITION APPARATUS, STORAGE MEDIUM AND NAVIGATION APPARATUS}

1 is a schematic configuration diagram of a navigation device.

2 is a flowchart for explaining a speech recognition process 1. FIG.

3 is a flowchart for explaining a speech recognition process 2. FIG.

4 is a flowchart for explaining a speech recognition process 3. FIG.

5 is a flowchart for explaining speech recognition processing 4. FIG.

6 is a screen image.

7 is a screen image.

8 is a screen image.

9 is a graph showing a change in recognition rate by the number of inputs.

* Explanation of symbols for the main parts of the drawings

20: navigation device 21: position detector

21a: GPS receiver 21b: gyroscope

21c: distance sensor 21d: geomagnetic sensor

22: operation switch group 23a: remote control

23b: remote sensor 25: map data input

26: display unit 27: audio output unit

28: microphone 29: control unit

30: Data input / output related to speech recognition

31: In-car LAN communication unit

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech recognition device or the like for determining a single syllable intended by a speaker based on a voice input by a speaker.

Background of the Invention A speech recognition apparatus that determines single syllables intended by a speaker by one syllable based on a speech input by a speaker is widely known. This type of speech recognition apparatus needs to be provided with a word dictionary corresponding to all the words for speech recognition in advance, such as a speech recognition apparatus for speech recognition in units of words (words composed of a plurality of single syllables). There is no advantage in that virtually anything can be recognized as the final set of recognition results (eg, sentences).

However, in the case of recognizing single-syllable speech, the recognition rate is generally low because recognition clues are smaller than speech recognition in word units. Therefore, in the speech recognition apparatus which recognizes such single-syllable speech, various studies have been conducted to improve the recognition accuracy. For example, the speaker may improve the recognition accuracy by studying and inputting a speech method, or confirm the speaker by outputting (talk back) the single syllable recognized by the speech recognition device to improve the final recognition accuracy. have.

 Here, the former method is demonstrated. In the speech recognition apparatus described in Patent Literature 1, a speaker recognizes a single syllable speech “A” by inputting, for example, “Aoi Eonoa”. As described above, by inputting a specific word for recognizing a single syllable longer than a single syllable, the speaker can improve the recognition accuracy of the speech recognition device as compared with the case of simply inputting a single syllable.

Patent Document 1: Japanese Patent Application Laid-Open No. 11-184495

However, it is true that even in the speech recognition apparatus using such a speech recognition method, it is difficult to completely prevent misrecognition due to the speaker's tone (so-called habit) or noise environment during speech. In addition, in the case of a speech recognition device that recognizes a single syllable voice, the speaker needs to correct or confirm every syllable syllable, and if there is a misperception, the speaker needs to work again.

This invention is made | formed in view of such a problem, and an object of this invention is to provide a speech recognition apparatus etc. which are as easy to use for a speaker as possible.

The speech recognition apparatus which concerns on one aspect of this invention made in order to solve the said subject contains a speech input means, a speech recognition means, a notification means, a receiving means, and a control means. The voice input means inputs the voice uttered by the speaker, the voice recognition means analyzes the voice input by the input means to specify candidate single syllables, the notification means notifies the designated information, and the receiving means accepts the speaker's operation. . Further, the control means performs a notification process for causing the speech recognition means to notify the notification means of the candidate most syllable having the highest potential among the specific candidate single syllables, and when the receiving means accepts an operation meaning decision from the speaker, To confirm the candidate single syllable notified at the time of the notification processing as a single syllable intended by the speaker, and when a new voice is input from the speaker to the voice input means and the speech recognition means specifies the candidate single syllable, the notification processing is performed. In return, the notifying candidate single syllable among the candidate single syllables is notified to the notifying means. In addition, the candidate single syllable mentioned here is literally a candidate of a single syllable, and one or more candidate single syllables which a speech recognition means specifies may be sufficient.

According to the speech recognition apparatus according to the above aspect, the speaker performs an operation only when the uttered single syllable is correctly recognized, and if it is not correctly recognized, the speaker can continue to utter the single syllable until it is correctly recognized without any manipulation. have. For this reason, when the recognition is not performed correctly, the speaker only needs to continuously re-ignite without performing the re-input command many times. In other words, it is convenient to use.

Moreover, the speech recognition apparatus which concerns on another aspect of this invention contains a speech input means, a speech recognition means, a notification means, and a control means. The voice input means inputs the voice uttered by the speaker, the voice recognition means analyzes the voice input by the input means to identify candidate single syllables, and also recognizes a definite word meaning definiteness, and the notification means notifies the designated information. do. In addition, the control means performs a notification process of notifying the notification means of the candidate single syllable having the highest potential among the specific candidate single syllables by the voice recognition means, and a new voice is input from the speaker to the voice input means so that the voice recognition means recognizes the final word. In the case where the candidate single syllable notified at the time of the previous notification processing is determined as the single syllable intended by the speaker, and when a new voice is input from the speaker to the voice input means and the speech recognition means specifies the candidate single syllable, Then, the processing returns to the execution of the notification process, and the notification means notifies the candidate single syllable having the highest potential among the candidate single syllables. Note that the candidate single syllable referred to herein is a literal candidate and the candidate single syllable specified by the speech recognition means may be one or plural.

According to the speech recognition apparatus according to another aspect of the present invention, the speaker is a definite word (e.g., "Next") or "Next" only when the spoken single syllable is correctly recognized. E)) '' or `` Next, '' to determine the single syllable by uttering it, and if it is not recognized correctly, to recognize it until it is recognized without any special operation or utterance. You can continue to sing single syllables. For this reason, if the speaker does not recognize correctly, the speaker only needs to re-ignite continuously without giving re-instruction many times. That is convenient to use.

However, if the recognized single syllable is not correct, there is a possibility that an inappropriate candidate single syllable is notified again, just as before, even when the speaker re-ignites. In order to prevent this, as described in another aspect of the present invention, when the control means performs the notification process two or more times in succession without performing the confirmation process, the candidate for notifying the candidate single syllable notified in the past by the notification process. Except from the single syllable, the most probable candidate single syllable may be notified.

In this way, the speaker may not be notified of the same inappropriate candidate single syllable again at the time of re-ignition, thereby improving the speaker's ease of use.

However, even if the correct candidate single syllable is notified, a case of accidental re-ignition may be considered. This mistake causes the problem that the correct candidate single syllable cannot be notified again. In order to prevent such a problem from occurring, according to still another aspect of the present invention, when there is a predetermined number of re-ignitions, the exclusion of candidate single syllables may be canceled. That is, if the control means does not exclude the candidate single syllable notified by the notification processing performed a predetermined number of times except for the immediately preceding one of the notification processing repeatedly performed without performing the confirmation processing for the above-mentioned exclusion. good.

The optimum value of the predetermined number of times may be three times in accordance with another aspect of the present invention. Based on the experiment conducted by the inventors of the present invention (the detailed description of this experiment is described in the paragraph of the embodiment), the probability of notifying the correct candidate single syllable up to four times of utterance is 98%. Even if repeated, the correct candidate single syllables are rarely notified after that. That is, in most cases, it means that the correct single syllable is notified once until three times the number of re-ignitions. If the number of the re-ignitions is three times, it means that the speaker is likely to have excluded the correct candidate single syllable by mistake.

Therefore, according to another aspect of the present invention, if the candidate single syllable notified by the notification processing performed before the past three times is not excluded, as described above, the problem that the correct candidate single syllable cannot be notified again is solved. It can prevent.

As described above, the predetermined number of times is optimally three times in experiments, but it is also possible to change it in rare cases due to factors such as the environment in which the speech recognition apparatus is used or the manner of speaking of the speaker. . For this reason, according to another aspect of the present invention, the control means may be configured to change the predetermined number of times based on the operation of the speaker accepted by the receiving means. If this is the case, the speaker can change the predetermined number of times according to the environment in which the speech recognition apparatus is used or the manner of speaking of the speaker.

Moreover, the speech recognition apparatus which concerns on another aspect of this invention contains a speech input means, a receiving means, a memory means, and a speech recognition means. The voice input means inputs the voice uttered by the speaker, the reception means receives the speaker's operation, and the storage means specifies a dictionary composed of a single syllable recognition word consisting of a plurality of syllables corresponding to each single syllable. Separate and remember multiple words. Further, the speech recognition means selects the dictionary stored by the storage means based on the operation of the speaker accepted by the reception means, and selects a single syllable recognition word corresponding to the speech input by the speech input means from the selected dictionary. And selects a single syllable corresponding to the selected single syllable recognition specific word as the syllable intended by the speaker.

Conventionally, in a voice of a specific word for speech recognition of a syllable composed of a plurality of syllables such as `` aiueeo no あ (Aiueo's) '' and `` asahino あ (Asahi's A) ''. There was a speech recognition device that recognized single syllables. However, in such an input speech recognition apparatus, the speaker needs to remember all the single syllable recognition specific words, and it is difficult to remember the single syllable recognition words that are not interesting from the speaker's point of view.

Here, if the speech recognition apparatus according to another aspect of the present invention stores a dictionary of single syllable recognition words divided into various genres or the like in advance, the speaker may select a specific syllable recognition word for one syllable that he or she easily remembers from among the dictionaries. The dictionary can be changed by operation. In this way, if the dictionary according to the speaker's taste can be used, the speaker can quickly store a specific word for single syllable recognition, and can use the speech recognition device skillfully.

In addition, according to still another aspect of the present invention, such a dictionary syllable recognition specific word may be registered by a speaker in advance by operating a receiving means. If this is the case, since a specific syllable for recognizing single syllables that reflects the taste of the speaker can be used, the voice recognition device can be used skillfully.

In addition, the speech recognition apparatus according to another aspect of the present invention is a speech recognition apparatus for determining a single syllable intended by the speaker based on the speech input by the speaker, the speech input means for inputting the speech spoken by the speaker; And speech recognition means for dividing the repetitive voice composed of the same single syllables input by the voice input means into voices for each single syllable, and determining the single syllable intended by the speaker based on the respective voices.

In this way, for example, when a speaker speaks "Aa Aa" and a speaker speaks, a single syllable called "A A" is recognized. As a result, the recognition clue is increased compared to the case where the speaker speaks `` A '', so the recognition rate is improved, and the burden of the speaker is also reduced because there is no need to store a specific word for single syllable recognition.

In addition, the speech recognition apparatus according to the present invention is a speech recognition apparatus for determining a single syllable intended by a speaker based on the speech input by the speaker, and includes speech input means for inputting speech spoken by the speaker. . In addition, when the voice of the single syllable input by the voice input means is any of a sound of a sound, a sound of a sound, a sound of a sound, and a sound of a half sound, the sound corresponding to the sound of sound, sound of a sound, sound of a sound, and sound of a sound If the voice is determined as a single syllable intended by the speaker, and the voice input by the voice input means is a specific word meaning a predetermined sound, the previously determined single syllable is changed to a corresponding single syllable of the corresponding sound. In the case where the input voice is a specific word meaning a predetermined note, the short syllable determined immediately before is changed to a short syllable of the corresponding note, and when the voice input means is a specific word meaning a predetermined tactile sound, If the determined single syllable is changed to a single syllable of the corresponding tactile sound, and the voice input by the voice input means is a specific word meaning a predetermined halftone sound, It includes speech recognition means for changing a monosyllabic a monosyllable of bantak sound corresponding. In addition, "hearing" referred to here means 45 (normal) basic single-syllable groups except a sound, a sound, a sound, and a half sound.

In general, it is more difficult to recognize the sound and non-sound in a single syllable than in recognizing different single syllables. Therefore, the recognition rate improves if it is supposed to recognize the sound and non-sound as a group and later change the sound to sound. To change later, for example, when a voice is input by the speaker as "て ん て ん (텡텡)", the single syllable input immediately before is changed into a sound tone. The same applies to the note sound, the touch sound, and the halftone sound.

In addition, a speech recognition apparatus according to another aspect of the present invention is a speech recognition apparatus for determining a single syllable intended by a speaker based on a speech input by a speaker, and includes: speech input means for inputting a speech spoken by a speaker; A memory means for storing a dictionary corresponding to a syllable syllable and a speech input by a voice input means for analyzing a combination of a syllable-recognition-specific word composed of syllables, specifying a combination of a specific syllable for a syllable syllable, and a syllable corresponding to a specific combination And speech recognition means for determining the result based on a dictionary stored by the storage means. For example, if the speaker utters "Kei" (K) and "Ai" (A), it will be recognized as "Ka", and "Kei" will be described. (K) and "Ai" (I) are speech recognition devices that recognize as "ki" when the speaker speaks. In addition, if the speaker utters `` I チ '' (1) and `` I チ '' (1) in correspondence with the row number and column number of the 50th note (which is the 50th note in Japanese), "あ (아)" It may be recognized as ".

Such a speech recognition apparatus improves the recognition rate because the speech length and the number of speeches of the recognition target are increased. In addition, since it is not necessary to prepare a specific word for recognizing single syllables for all single syllables (as in the example described above, "ケ (kei)" can be used to recognize all single syllables in the line ka), The capacity is reduced, and the single syllable recognition word that the speaker also stores is reduced, thereby improving convenience of use.

Moreover, the speech recognition apparatus which concerns on another aspect of this invention contains a speech input means, a speech recognition means, a display means, and a control means. The voice input means inputs the voice uttered by the speaker, the voice recognition means analyzes the voice input by the voice input means to specify candidate single syllables, and the display means has a display area for displaying specified information, It has a sensor that senses the speaker's manipulation to the surface along with its position in the display area. In addition, when the speech recognition means specifies a plurality of candidate single syllables, the control means lists and displays each object corresponding to the candidate single syllables in the display area of the display means as the largest group of objects in the display area. The candidate single syllable corresponding to the object displayed at the position detected by the sensor is determined as the single syllable intended by the speaker.

In this case, the speaker can visually confirm the candidate single syllables, and can display the candidate single syllables by day because the speakers are listed and displayed as the largest object group in the display area. As a result, the speaker can smoothly determine the single syllable.

However, the number of objects corresponding to the candidate single syllables to be displayed on the display means may be any number, but according to another aspect of the present invention, only three are selected in ascending order according to the prospect of candidate single syllables corresponding to the object. It is good to display in. This is because increasing the number of objects to be displayed increases the probability that an accurate single syllable is determined by one utterance. However, if too many objects are displayed, the list is deteriorated at this time, which makes it difficult for the speaker to select objects. Therefore, considering the probability that any one of the objects to maintain this balance and display contains the syllables intended by the speaker, three objects are most suitable.

The objects displayed in the display area of the display means may all have the same size, color, or shape, but according to another aspect of the present invention, the display means may be changed by changing the visual characteristics in accordance with the prospect of candidate single syllables corresponding to the object. The term " modifying visual characteristics " may specifically consider size, color, shape, blinking state, animation, and the like.

In this way, for example, when three candidate single syllables are displayed, it is possible to know which candidate single syllables are most promising by day, and the speaker can easily select candidate single syllables. In addition, when such a voice recognition device is used in a vehicle, for example, the time for the driver to watch the display means can be shortened.

Further, according to another aspect of the present invention, the control means handles a specific range that specifies each object by the position detected by the sensor, rather than the display range occupied by each object in the display area of the display means at the time of determination. It is good to be.

In this way, the speaker does not need to touch the position where the object is displayed accurately. Thus, for example, when the voice recognition device is mounted on a vehicle, the driver checks the display means and then selects the desired candidate syllable even if the candidate is selected by guessing the place without actually looking at the display means when actually operating. It is more likely to be.

Moreover, the speech recognition apparatus which concerns on another aspect of this invention contains a speech input means, a speech recognition means, a display means, and a control means. The voice input means inputs the voice uttered by the speaker, the voice recognition means analyzes the voice input by the voice input means to specify candidate single syllables, and the display means has a display area for displaying specified information, It has a sensor that senses the speaker's manipulation to the surface along with its position in the display area. Further, the control means displays an object representing a single syllable in correspondence with 50 notes in the display area of the display means, and when the speech recognition means specifies a plurality of candidate single syllables, Visual features are displayed differently from other objects, and the candidate single syllables represented by the objects displayed at the position detected by the sensor by the speaker's operation are determined as the single syllables intended by the speaker. .

With such a speech recognition device, the speaker can determine the single syllable from the candidate single syllables recognized by the speech recognition device, and if the intended single syllable does not exist among the candidate single syllables, the speaker can directly specify the single syllable among the 50 notes.

Moreover, the speech recognition apparatus which concerns on another aspect of this invention is a speech input means, a speech recognition means, a display means, a vehicle state acquisition means, and the control means of the speech recognition apparatus in any one aspect of this invention, Control means of the speech recognition apparatus, and main control means. The voice input means inputs the voice uttered by the speaker, the voice recognition means analyzes the voice input by the voice input means to specify candidate single syllables, and the display means has a display area for displaying the specified information, A sensor for detecting the operation of the speaker with the position in the display area, and the vehicle state obtaining means acquires information of whether the vehicle is running. Further, the main control means functions the control means of the speech recognition apparatus according to one aspect of the present invention when it is determined that the vehicle is running based on the information acquired by the vehicle state obtaining means, and when it is determined that the vehicle is stationary, the present invention. Function control means of the speech recognition apparatus according to one aspect of the present invention.

In this way, the single-syllable can be determined in a manner that does not require the viewer to keep an eye on the display means when the speaker is in operation. You can enter a single syllable and confirm it.

In addition, the speech recognition apparatus according to the aspect of the present invention includes speech input means, speech recognition means, notification means, storage means, and control means. The voice input means inputs the voice uttered by the speaker, the voice recognition means analyzes the voice input by the voice input means to identify the candidate single syllables, the notification means notifies the designated information, and the storage means corresponds to a plurality of single syllables. It stores a dictionary composed of confirmation words made up of syllables. Furthermore, the voice recognition means performs a notification process for causing the notification means to notify the notification means by means of recognition words stored in the storage means by the speech recognition means having the most likely candidate single syllable among the specific candidate single syllables, and the receiving means receives an operation meaning a decision from the speaker. If accepted, the candidate single syllable to be notified is determined as the speaker's intended single syllable.

In this way, it is possible to accurately identify single syllables that are difficult for the speaker to recognize.

However, when notifying a word for confirmation instead of a single syllable in this way, it is sufficient to be able to use the word for confirmation according to the taste of the speaker. That is, as in the speech recognition apparatus according to the aspect of the present invention, the storage means divides the dictionary according to the type of the word for recognition, and stores the plural numbers, and the control means stores the memory means based on the operation of the speaker accepted by the reception means. The dictionary to be selected may be selected, and notification processing may be performed using the selected dictionary.

If so, the speaker can identify the syllable by the preferred confirmation word.

In addition, according to another aspect of the present invention, the confirmation word may be registered so that the speaker can operate in advance by operating the receiving means. If so, the confirmation word reflecting the speaker's taste can be notified.

In addition, even when the candidate single syllable is displayed on the display means, the identification words corresponding to the candidate single syllables may be displayed instead of the candidate single syllables. That is, according to another aspect of the present invention, there is provided a display means having a display area for displaying additionally designated information and having a sensor for detecting a speaker's manipulation of the surface of the display area together with a position in the display area. When the speech recognition means specifies a plurality of candidate single syllables, the control means lists the identification words corresponding to the candidate single syllables as objects on the display means, displays them as objects, and is operated by the speaker to detect the sensors. The candidate single syllable corresponding to the object displayed at one position may be determined as the single syllable intended by the speaker.

In addition, the confirmation word displayed in this manner may be configured to be confirmed by the speaker speaking. That is, according to another aspect of the present invention, the control means, after the speech recognition means specifies a plurality of candidate single syllables, displays confirmation words corresponding to the candidate single syllables as objects on the display means, and then the speech recognition means The candidate single syllable corresponding to the specific confirmation word may be determined as the single syllable intended by the speaker.

In this way, the speaker can confirm the candidate single syllable by speech, and can also firmly determine the candidate single syllable.

However, according to another aspect of the present invention, a computer in which the speech recognition apparatus is built may be executed by a program which functions as at least one of the speech confirmation means and the control means in the speech recognition apparatus. If so, for example, the program is recorded on a computer-readable recording medium such as a flexible disk, a magneto-optical disk, a CD-ROM, a hard disk, a ROM, or a RAM, and the program is loaded into the computer as necessary. By activating, the mobile phone can function as at least one of a voice recognition means and a control means in the voice recognition device. In addition, since the program can also be distributed using a network or the like, the function of the speech recognition device can be easily improved.

Moreover, according to another aspect of the present invention, the same applies to a program for functioning as a voice confirmation means in the voice recognition device. In addition, according to an aspect of the present invention, the same applies to a program functioning as a main control means in the speech recognition apparatus. Of course, such a program may be stored in the storage medium according to the aspect of the present invention.

However, the speech recognition apparatus may be associated with the navigation apparatus according to the aspect of the present invention, and the single speech syllable group acquired by the speech recognition apparatus may be used when the navigation apparatus performs the navigation process. The navigation process referred to here is, for example, a process of displaying a map and displaying a current place on the map, or a route guide process of performing route guidance according to a set route.

In this way, various operations performed by the user in the navigation process can be performed by voice, thereby improving the convenience of use of the navigation process.

EMBODIMENT OF THE INVENTION Hereinafter, the Example to which this invention was applied is demonstrated using drawing. In addition, the embodiment of the present invention is not limited to the following examples, and may take various forms as long as it falls within the technical scope of the present invention.

(First embodiment)

1 is a block diagram showing the configuration of a navigation device 20 having a voice recognition function. The navigation device 20 is mounted on a vehicle, and is the same as the position detector 21 for detecting the current position of the vehicle, the operation switch group 22 for inputting various instructions from the user, and the operation switch group 22. Instructions can be input, the remote control terminal (hereinafter referred to as "remote control") 23a separate from the navigation device 20, the remote control sensor 23b for inputting signals from the remote control 23a, and map data. Or a map data input unit 25 for inputting map data or the like from a map storage medium on which various kinds of information are recorded, a display unit 26 for performing a map or various information display, and an audio output unit 27 for outputting various guide voices and the like. A microphone 28 for inputting voice and outputting voice information, a voice recognition data input / output device 30 for inputting and outputting voice recognition related data, an in-vehicle LAN communication unit 31 for communicating with the in-vehicle LAN, and From the above-described position detector 21, the operation switch group 22, the remote control sensor 23b, the map data input device 25, the microphone 28, the voice recognition data input / output device 30, and the in-vehicle LAN communication unit 31 The control unit 29 performs various processes according to the input and controls the display unit 26, the voice output unit 27, the voice recognition related data input / output unit 30, and the in-vehicle LAN communication unit 31.

The position detector 21 receives a radio wave from a satellite for a GPS (Global Positioning System) through a GPS antenna, and applies a GPS receiver 21a for detecting the position, orientation, and speed of the vehicle to the vehicle. A gyroscope 21b for detecting the magnitude of the rotational movement, a distance sensor 21c for detecting the distance traveled from the acceleration in the front-rear direction, etc. of the vehicle, and a geomagnetic sensor 21d for detecting the traveling direction from the geomagnetic Doing. Each of these sensors 21a to 21d is configured to be used while complementing each other because they have different errors in their properties.

The operation switch group 22 is composed of a touch panel integrally formed with the display surface of the display unit 26 and a mechanical key switch or the like provided around the display unit 26. The touch panel and the display unit 26 are stacked and integrated, and there are various methods such as a pressure reducing method, an electromagnetic induction method, a capacitive method, or a combination thereof, but any one of them may be used.

The map data input unit 25 is a device for inputting various data stored in a map storage medium (not shown). The map storage medium stores map data (road data, terrain data, mark data, intersection data, facility data, etc.), voice data for guidance, voice recognition data, and the like. As a type of map storage medium that stores such data, a CD-ROM or a DVD is generally used, but a magnetic storage device such as a hard disk or a medium such as a memory card may be used.

The display unit 26 is a color display device, and there are liquid crystal displays, organic EL displays, CRTs, and the like, but any of them may be used. On the display screen of the display unit 26, a mark indicating a specific current place from a vehicle current position detected by the position detector 21, map data input from the map data input unit 25, a guidance route to a destination, a name, a sign And additional data such as marks of various facilities can be displayed in duplicate. It is also possible to display guides of facilities.

The audio output unit 27 may output a facility guide or various guide voices input from the map data input unit 25.

The microphone 28 outputs an electric signal (voice signal) to the control unit 29 based on the input voice when the user inputs (speaks) a voice. The user can operate the navigation device 20 by inputting various voices into the microphone 28.

The speech recognition related data input / output unit 30 is an apparatus for inputting and outputting various data stored in a speech recognition related data storage medium (not shown). The speech recognition-related data storage medium includes a dictionary composed of feature parameters for recognizing single syllables, a single syllable recognition word consisting of a plurality of syllables corresponding to each syllable, and a confirmation word composed of a plurality of syllables corresponding to each syllable. The dictionary is memorized. As a type of map storage medium storing such data, a magnetic storage device such as a hard disk or a medium such as a memory card may be used.

The in-vehicle LAN communication unit 31 is connected to the in-vehicle LAN and can communicate with various ECUs connected to the in-vehicle LAN. As the in-vehicle LAN, for example, CAN (Control Aria Network) is assumed, and one of the various ECUs is assumed to be an engine ECU, an AT-ECU, or a Bode-ECU.

The control circuit 29 is mainly comprised of a well-known microcomputer consisting of a CPU, a ROM, a RAM, an I / O, a bus line connecting these structures, and the like stored in a program stored in the ROM and the RAM. Various processing is performed based on this. For example, the current position of the vehicle is calculated based on each detection signal from the position detector 21 as a set of coordinates and the direction of travel, and a map near the current position read through the map data input unit 25 is obtained. Optimal from the current position to the destination based on the display processing displayed on the display unit 26 or the destination data set in accordance with the point data stored in the map data input unit 25 and the operation of the operation switch group 22 or the remote controller 23a. The route is calculated and route guidance processing for guiding the calculated route is performed. In addition, the controller 29 may perform a voice recognition process to be described later.

Although the schematic structure of the navigation apparatus 20 was demonstrated so far, the correspondence of each part of the navigation apparatus 20 and the term described in a claim is shown. The microphone 28 corresponds to the voice input means, the audio output portion 27 corresponds to the notification means, the display portion 26 corresponds to the notification means and the display means, and the operation switch group 22 and the remote controller 23a. The control unit 29 corresponds to the voice recognition unit, the control unit and the main control unit, the voice recognition related data storage medium corresponds to the storage unit, and the in-vehicle LAN communication unit 31 corresponds to the vehicle state acquisition unit. It is considerable.

Next, the voice recognition process 1 performed during inputting a name of a destination or the like before the route guidance process among the processes performed by the control unit 29 will be described using the flowchart of FIG. The speech recognition processing 1 starts to be performed when the user specifically instructs the user to input a voice at the time of inputting the information to the navigation apparatus 20.

When the control unit 29 starts performing, first, the process branches to whether or not the torque SW installed in the operation switch group 22 and the remote control unit 23a has been pressed by the user (S110). If the torque SW is pressed by the user, the processing proceeds to the next processing step, otherwise it is maintained at this step.

In following S115, a confirmation sound (for example, an electronic sound of "beep" or a guide voice of "Please input voice") is output to the voice output unit 27.

In the following S120, the user's voice is input through the microphone 28.

In subsequent S125, the voice input in S120 is analyzed (feature parameter extracted), and the candidate single syllable is given a plurality of candidates by selecting a candidate rank by comparing the characteristic parameters of the single syllable acquired through the voice recognition-related data input / output unit 30 and the like. .

In subsequent S130, candidate single syllables in the exclusion buffer are excluded from the candidate single syllables selected in S125. This exclusion buffer is a buffer existing in the control unit 29 and capable of storing three exclusion-designated candidate single syllables. In addition, the exclusion buffer is initialized at the start of execution of the voice recognition process 1.

In subsequent S135, the candidate single syllable having the highest candidate rank among the candidate single syllables is displayed on the display unit 26 or the audio output unit 27 is outputted as voice.

Next, whether the fixed SW (may be shared with the above-mentioned torque SW) provided in the operation switch group 22 and the remote controller 23a has been pressed by the user, or whether voice has been further input by the user. The process branches according to step S140. When the definite SW is pressed by the user, the flow advances to S145. When the definite SW is not operated by the user and further voice is input by the user, the flow proceeds to S150.

In S145, the candidate single syllable notified in S135 is confirmed as the final single syllable, and added to the end of the already determined final single syllable group. The exclusion buffer is also initialized (S153). When the exclusion buffer is initialized, the processing branches according to whether or not the end SW installed in the operation switch group 22 or the remote controller 23a has been operated by the user (S155). If the end SW is operated by the user, this process (voice recognition process 1) is terminated, and if the end SW is not operated by the user, the process returns to the above-described S115.

On the other hand, in S150, the candidate single syllable notified in S135 is input to the exclusion buffer. At this time, if there are already three candidate single syllables in the exclusion buffer, the candidate single syllables that have been previously input to the exclusion buffer are deleted, and the candidate single syllables newly notified in S135 are input as the exclusion buffer.

For convenience (to simplify the explanation), only the step S155 for determining whether or not the end SW has been operated is to terminate the voice recognition processing 1 by the operation of the end SW. In this case, the voice recognition process 1 is terminated immediately. In addition, in the voice input steps S120 and S140 or the user's operation wait step S140, the voice recognition process 1 is terminated even when there is no input of the voice for a predetermined time (for example, 30 seconds) or the user's operation. .

Although the voice recognition processing 1 has been described so far, the single-syllable group is used as the name of the destination in the route guidance process or as the name of the facility.

According to such a navigation device 20, the user performs an operation only when the uttered single syllable is correctly recognized to determine the single syllable, and if it is not correctly recognized, continues to perform the single syllable until it is correctly recognized without any manipulation. Can ignite. For this reason, the user only has to re-ignite continuously without giving re-input again and again if the recognition is not made correctly. That is convenient to use.

In addition, since the candidate single syllables stored in the exclusion buffer are excluded from the newly selected candidate single syllables by re-ignition, inappropriate candidate single syllables similar to the previous time are not notified again at the time of re- firing, and thus are convenient for the user to use.

The reason why the above-described exclusion buffer is configured to store only three candidate single syllables will be described.

The inventors of the present invention performed the following experiment. This experiment is an experiment in which two men and women of each age (ie, 20 persons in total) from 20 to 60 years of age in a stationary car are repeatedly fired 10 times, one by three. In addition, on the basis of this experimental result, a graph showing the probability of correcting a single syllable up to this input number on the vertical axis is shown in FIG. As can be seen from the graph, the recognition rate is almost constant after the third round (the third round is 96%, the fourth round is 98%, the fifth round is 98%), and after that, there is almost no change. In other words, even if the number of times of speech is duplicated four or more times, the correct candidate single syllable is rarely notified thereafter. That is, in most cases, it means that the correct single syllable is notified once until three times the number of re-ignitions, and when the number of re-ignitions is three times, it means that the speaker is likely to exclude the correct candidate single syllables by mistake. Therefore, when the number of times of re-ignition is three, it is better to be able to notify the recognition candidate once again that the first recognition has been made.

If this is the case, even though the candidate single syllable is notified, even if the user accidentally re-ignites, the correct candidate single syllable is not reported again because the excluded candidate single syllable is returned to the state obtained by being notified again at an appropriate timing. This can prevent the problem of being turned off.

(2nd Example)

Next, a second embodiment will be described. Since the navigation apparatus having the voice recognition function of the second embodiment has the same configuration as the navigation apparatus 20 of the first embodiment described above, only the differences will be described. The main difference lies in the speech recognition process performed by the control unit 29. Hereinafter, the speech recognition process 2 performed by the control unit 29 will be described using the flowchart of FIG.

The speech recognition processing 2 starts to be performed when the user specifically instructs in the state in which voice input is possible at the time of inputting information to the navigation apparatus 20.

When the control unit 29 starts performing, first, the process branches according to whether or not the torque SW installed in the operation switch group 22 or the remote control unit 23a is pressed by the user (S210). If the torque SW is pressed by the user, the processing proceeds to the next processing step, otherwise it is held in this step.

In subsequent S215, a confirmation sound (for example, an electronic sound of "beep" or a guide voice of "Please input voice") is output to the audio output unit 27.

In S220 that follows, the user's voice is input through the microphone 28.

In subsequent S225, the voice inputted in S220 is analyzed (feature parameter extracted), and the candidate single syllable is given a plurality of candidates by selecting a candidate rank by comparing with the feature parameter of the single syllable acquired through the voice recognition-related data input / output unit 30 and the like. . In addition, when the voice input in S220 is not a single syllable, it is determined whether it is a definite word ("next", "next", "next", etc.) that means confirmation).

In subsequent S230, the processing branches according to whether or not the voice input in S220 is a definite word meaning definite. If the voice input in S220 is a definite word, the flow proceeds to S250. If the voice input in S220 is not a definite word, the flow proceeds to S235.

In S235, the candidate single syllables in the exclusion buffer are excluded from the candidate single syllables selected in S225. This exclusion buffer is a buffer existing in the control unit 29 and capable of storing three exclusion-designated candidate single syllables. Also, the exclusion buffer is initialized at the start of the performance of the speech recognition process 2.

Further, in S240, the candidate single syllable having the highest candidate rank among the candidate single syllables is displayed on the display unit 26, or the audio output unit 27 outputs and notifies the voice.

In S245, the candidate single syllable notified in S240 is input to the exclusion buffer. At this time, if there are already three candidate single syllables in the exclusion buffer, the candidate single syllables inputted to the exclusion buffer most recently are deleted, and the candidate single syllables newly notified in S240 are input as the exclusion buffer.

On the other hand, in S230, in step S250 where the voice input in S220 proceeds as a definite word, the candidate single syllable notified last time is determined as the definite single syllable, and is added to the end of the definite single syllable group already determined. The exclusion buffer is initialized (S253). When the exclusion buffer is initialized, the processing branches according to whether or not the end SW installed in the operation switch group 22 and the remote controller 23a has been operated by the user (S255). If the end SW has been operated by the user, this process (voice recognition process 2) is terminated. If the end SW has not been operated by the user, the process returns to S215 described above.

Also, for convenience (to simplify the explanation), the voice recognition process 2 by the operation of the end SW is to be terminated only in step S255 in which it is determined whether the end SW has been operated. The voice recognition process 2 is immediately terminated. In addition, in the voice input step S220, the voice recognition process 2 is terminated even when there is no input of voice for a predetermined time (for example, 30 seconds).

Although the voice recognition processing 2 has been described so far, the single-syllable group determined in this manner is used as the name of the destination or the name of the facility during the route guidance processing.

According to such a navigation 20, the user utters a definite word ("next") only when the uttered single syllable is correctly recognized, and if it is not correctly recognized, without any special manipulation or utterance, You can continue to utter the single syllable you want to recognize until it is correctly recognized. For this reason, the user only needs to re-ignite continuously without giving a re-entry instruction continuously if the recognition is not made correctly. That is convenient to use.

(Third Embodiment)

Next, a third embodiment will be described. Since the navigation device having the voice recognition function of the third embodiment has the same configuration as the navigation device 20 of the first embodiment described above, only the differences will be described. The main difference lies in the speech recognition process performed by the control unit 29. Hereinafter, the speech recognition process 3 performed by the control unit 29 will be described using the flowchart of FIG.

When the control unit 29 starts performing, first, the process branches to whether or not the torque SW installed in the operation switch group 22 and the remote control unit 23a has been pressed by the user (S310). If the torque SW is pressed by the user, the processing proceeds to the next processing step, otherwise it is maintained at this step.

In subsequent S315, a confirmation sound (for example, an electronic sound of "beep" or a guide voice of "Please input voice") is output to the audio output unit 27.

In the following S320, the user's voice is input through the microphone 28.

In subsequent S325, the voice input in S320 is analyzed (feature parameter extracted), and three candidate single syllables are selected by comparing the characteristic parameters of the single syllable acquired through the speech recognition related data input / output unit 30 and the like.

In subsequent S330, vehicle speed information is acquired from an engine ECU (not shown) via the in-vehicle LAN communication unit 31, and the process branches according to whether the vehicle is running. If the vehicle is running, the process proceeds to S335. If the vehicle is not driving, the process proceeds to S340.

In S335, the candidate single syllables selected in S325 are displayed on the display unit 26, listed as the largest object group in the display area. An example of this display is shown in FIG. As shown in Fig. 6, the screen 100 displays the candidate single syllable objects 101 to 103 arranged in such a manner as to occupy most of the display area. In addition, an operation specific range 104 indicated by a dotted line (not actually displayed, which is the same below) is set in an area wider than the candidate single syllable object 101. This operation specifying range 104 is a range in which the control unit 29 recognizes that the candidate single syllable object 101 has been selected by the user when the user touches the operation specifying range 104. In this way, the operation specific range 105 is set in the candidate single syllable object 102, and the operation specific range 106 is set in the candidate single syllable object 103.

4, on the other hand, in S340, 50 notes are displayed on the display unit 26, and the border of the object of the candidate single syllable selected in S325 is changed. An example of this display is shown in FIG. As shown in Fig. 7, each single syllable is arranged as an object in the form of a list of 50 sounds on the screen 111, and among them, candidate objects of "A", "Ha" and "Ma". Only 112 to 114 have a thickness and a color of a border different from other single syllable objects.

Returning to Fig. 4, in S345 which follows, the processing branches according to whether an object is selected by the user based on the signal output from the touch panel integrally formed with the surface of the display unit 26. If an object is selected by the user, the process proceeds to S350. If no object is selected by the user (for example, 30 seconds), the process returns to the above-described S320.

In S350, which proceeds when an object is selected by the user, the candidate single syllable corresponding to the selected object is determined as the determined single syllable and added to the end of the determined single syllable group. Note that the term " selected object " referred to herein is not limited to the candidate single syllable object when the display described in S340 (see FIG. 7) is performed, and any object of the single syllable selected by the user is a target. .

In subsequent S355, the processing branches according to whether or not the end SW provided in the operation switch group 22 or the remote controller 23a has been operated by the user. If the end SW is operated by the user, this process (voice recognition process 3) is terminated, and if the end SW is not operated by the user, the process returns to the above-described S315.

Further, only in step S355 for determining whether or not the end SW has been operated for convenience (to simplify the explanation), the voice recognition process 3 by the operation of the end SW is to be terminated. In this case, the voice recognition process 3 is terminated immediately. In the voice input step S320, the voice recognition process 3 is terminated even when there is no voice input for a predetermined time (for example, 30 seconds).

Although the voice recognition processing 3 has been described so far, the single-syllable group determined in this way is used as the name of the destination during the route guidance process or as the name of the facility.

According to such a navigation device 20, when the vehicle is driving, the candidate single syllables are listed and displayed as the largest object group in the display area of the display unit 26, so that the user can check the single syllables one by one. As a result, the user can smoothly determine the single syllable. In this case, since the specific range for specifying each object is wider than the display range occupied by each object in the display area of the display unit 26, the user can accurately determine the position at which the object is displayed. There is no need to touch. Therefore, even if the user is driving, the user can easily select the desired candidate single syllable.

On the other hand, when the vehicle is stopped, the user can select single syllables other than the candidate single syllables, so that the single syllable can be determined more quickly.

(Example 4)

Next, a fourth embodiment will be described. Since the navigation apparatus having the voice recognition function of the fourth embodiment has the same configuration as the navigation apparatus 20 of the first embodiment described above, only the differences will be described. The main difference lies in the speech recognition process executed by the control unit 29. Hereinafter, the speech recognition process 4 executed by the control unit 29 will be described using the flowchart of FIG.

When the control unit 29 starts performing, first, the process branches to whether or not the torque SW installed in the operation switch group 22 and the remote controller 23a has been pressed by the user (S410). If the torque SW is pressed by the user, the processing proceeds to the next processing step, otherwise it is held in this step.

In following S415, a confirmation sound (for example, an electronic sound of "beep" or a guide voice of "Please input voice") is output to the audio output unit 27.

In S420 that follows, the user's voice is input through the microphone 28.

In subsequent S425, the speech inputted in S320 is analyzed (feature parameter extracted, etc.), and three candidate single syllables are selected by comparing the characteristic parameters of the single syllable acquired through the speech recognition related data input / output unit 30 and the like.

In S435, the confirmation word corresponding to the candidate single syllable selected in S425 is displayed and displayed as a group of objects in the display area of the display unit 26, and the voice output unit 27 is notified sequentially as voice. The confirmation word mentioned here is a word which can be acquired through the speech recognition data input / output device 30, and includes the single syllable at the head corresponding to each single syllable. Specifically, for example, the confirmation word "Asahi" and the syllable "Ha (ha)" for the single syllable "あ" and the single syllable "ま" ), "Matsuri" and the like. An example of this display is shown in FIG. As shown in FIG. 8, on the screen 121, confirmation word objects 122, 123, and 124 are arranged so as to occupy most of the display area. In addition, when a user touches any of the confirmation word objects 122-124, the control part 29 is able to recognize which touched confirmation word object.

5, in S440, the user's voice is input through the microphone 28. Then, the voice input in S440 is analyzed (feature parameter extracted, etc.), and at step S435, one of the confirmation words displayed on the display unit 26 is attempted (S445).

In S450 which follows, if it is possible to specify which of the confirmation words displayed on the display part 26 in S435, it progresses to S455, and when it cannot specify, the process returns to S420.

In S455, the candidate single syllable corresponding to the identifying word which can be identified is added as the final single syllable group as the final single syllable.

In subsequent S460, the processing branches according to whether or not the end SW provided in the operation switch group 22 and the remote controller 23a has been operated by the user. If the end SW has been operated by the user, this process (voice recognition process 4) is terminated. If the end SW has not been operated by the user, the process returns to S435 described above.

Also, for convenience (to simplify the explanation), the voice recognition processing 4 by the operation of the end SW is to be terminated only in step S460 for determining whether the end SW has been operated, but the end SW has been operated in any step. The voice recognition process 3 is immediately terminated. In the voice input steps S420 and S440, the voice recognition process 4 is terminated even when there is no voice input for a predetermined time (for example, 30 seconds).

Although the speech recognition processing 4 has been described so far, the definite single-syllable group determined in this way is used as the name of the destination or the name of the facility during the route guidance processing.

According to such a navigation device 20, since the candidate single syllable is notified to the user by using the words for confirmation, the user is easier to grasp than to be notified by the single syllable. In addition, even when a voice is selected from the candidates, since the word can be selected using the confirmation word, the recognition rate is high even when the selection is performed by voice.

Hereinafter, another Example is described.

(1) In the above embodiment, the user basically performs audio input in single syllables, but may be input by a single syllable recognition word consisting of a plurality of syllables corresponding to the single syllable. In this case, the navigation device 20 may be configured to specify a single syllable corresponding to the input single syllable recognition specific word based on the voice recognition related data input through the voice recognition related data input / output unit 30. In addition, if a dictionary of a single syllable recognition word divided into various genres is stored in a speech recognition data storage medium in advance and the user can select the dictionary, the dictionary can be selected according to the user's preference. The user can quickly memorize and use a specific word for single syllable recognition. If the user can register the single syllable recognition specific word, the user can quickly store and use the single syllable recognition specific word.

(2) In addition, as a method for analyzing a voice, the navigation device 20 divides a repeated voice composed of the same single syllable input into voices for each single syllable, and generates a single syllable intended by the user based on the respective voices. It may be decided. That is, when a user utters a single syllable consecutively (for example, "Aaa"), a single syllable "A (ah)" is recognized. In this case, the recognition rate is also improved because the recognition clue increases as compared with the case where the user speaks "A".

(3) In addition, the navigation device 20 is a method for analyzing a voice. When the input single-syllable voice is any of the sound of the sound, the sound of the sound, the sound of the sound, the sound of the sound, the sound of the sound, the sound of the sound. May be determined as a single syllable intended by the user. In addition, in that case, if the inputted voice is a specific word meaning, for example, a predetermined sound, it is sufficient to change the single syllable determined immediately before to the single sound syllable of the corresponding sound. In addition, as long as it is a specific word meaning a predetermined sound, the short syllable determined immediately before may be changed into the short syllable of the corresponding sound. The same applies to the tactile sound and the halftone sound. On the other hand, the term "hearing" as referred to herein means 45 basic monosyllable groups except for sound of sound, sound of sound, sound of sound, and sound of sound.

In general, it is more difficult to recognize the sound of a single syllable and the sound of a non-syllable than to recognize different single syllables. Therefore, the recognition rate improves if it is supposed to recognize that the sound is not the sound of the sound and the sound as a bundle and later change to the sound of sound of the sound. To change later, for example, when a voice is input by the user as "て ん て ん", it is sufficient to change the short-syllable input immediately before to sound. The same applies to the note sound, the touch sound, and the halftone sound.

(4) In addition, as a method of analyzing a voice, the navigation device 20 may be determined as a single syllable intended by the user based on a combination of single-syllable recognition specific words corresponding to the input Roman oral voice. For example, if a user inputs "Kay" (K) and "Aay" (A), it is recognized as "Ka", and "Kay" (K), " It is a navigation device that recognizes "Ki" when a user speaks with "AI" (I). If the speaker speaks "Ich" (1) and "Ich" (1) in correspondence with the row and column numbers of the 50th note, it may be recognized as "あ".

Such a speech recognition apparatus improves the recognition rate because the speech length and the number of speeches of the recognition target are increased. In addition, since it is not necessary to prepare a specific word for single syllable recognition for the whole syllable (as in the above example, `` KE (K) '' can be used to recognize the entire syllable of the カ line). The capacity of the dictionary is reduced, and the single syllable recognition word that the user also remembers is reduced, which further improves usability.

(5) In addition, when a voice is input, the navigation device 20 terminates the voice recognition process when the word indicates the end of the voice recognition process (for example, "end", "complete", etc.). It is good to be. In this case, the user can finish the speech recognition process even by speaking, and thus the usability is improved.

(6) In the fourth embodiment, the confirmation word is notified instead of the candidate single syllable itself, but the dictionary of the confirmation word divided by various genres or the like is stored in the speech recognition data storage medium in advance. If the user can select the dictionary, the dictionary can be selected according to the user's taste, so that the user can use the word for confirmation desired by the user. In addition, if the user can additionally register the confirmation word, the user can further use the confirmation word desired by the user.

According to the present invention, it is possible to provide a speech recognition device or the like which is as easy to use for the speaker as possible.

Claims (29)

In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; Speech recognition means for specifying a candidate single syllable by analyzing the speech inputted by the input means; Notifying means for notifying specified information; Receiving means for receiving a speaker's operation; And If the speech recognition means executes a notification process for notifying the notification means of the most probable candidate single syllable among the specific candidate single syllables, and the reception means accepts an operation meaning a decision from the speaker, the notification processing immediately before A confirmation process for determining the candidate single syllable notified by the speaker as a single syllable intended by the speaker, and when a new voice is input from the speaker to the voice input means and the voice recognition means specifies a candidate single syllable, execution of the notification process is performed. Control means for returning to notify said notifying means of the most probable candidate single syllable among the candidate single syllables Speech recognition device comprising a. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; Speech recognition means for analyzing a speech inputted by the input means to specify a candidate single syllable and to recognize a definite word meaning definite; Notifying means for notifying specified information; And The speech recognition means executes a notification process of notifying the notification means of the candidate single syllable having the highest potential among specific candidate syllables, and a new voice is input from the speaker to the speech input means so that the speech recognition means recognizes the final word. In one case, a determination process is performed to determine the candidate single syllable to be notified at the time of the previous notification processing as the single syllable intended by the speaker, and a new voice is input from the talker to the voice input means so that the voice recognition means replaces the candidate single syllable. In a specific case, the control means returns to the execution of the notification process and notifies the notification means of the candidate single syllable having the highest potential among the candidate single syllables. Speech recognition device comprising a. The method according to claim 1 or 2, When the control means executes the notification process two or more times in succession without executing the determination process, the most probable one is excluded from candidate single syllables that notify the candidate single syllables notified in the past by the notification process. To notify candidate syllables  Speech recognition device. The method of claim 3, The control means does not exclude candidate single syllables notified by the notification process executed a predetermined number of times except for the immediately preceding one of the notification processes repeatedly executed without executing the confirmation process with respect to the exclusion. Speech recognition device. The method of claim 3, The control means does not exclude candidate single syllables notified by the notification process that has been executed three times in the past among the notification processes repeatedly executed without executing the determination process with respect to the exclusion. Speech recognition device. The method of claim 4, wherein The control means changes the predetermined number of times based on the operation of the speaker accepted by the receiving means. Speech recognition device. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; Receiving means for receiving a speaker's operation; Storage means for classifying and storing a dictionary composed of a single syllable recognition specific word consisting of a plurality of syllables corresponding to each single syllable for each kind of the single syllable recognition specific word; And Based on the operation of the speaker accepted by the receiving means, the dictionary stored in the storage means is selected, and from the selected dictionary, a specific word for single syllable recognition corresponding to the voice input by the voice input means is selected, and the selected dictionary is selected. Speech recognition means for determining a single syllable corresponding to the specific syllable for single syllable recognition as a single syllable intended by a speaker Speech recognition device comprising a. The method of claim 7, wherein The specific syllable recognition word of the dictionary stored in the storage means can be registered by a speaker operating the receiving means in advance. Speech recognition device. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; And Speech recognition means for dividing a repetitive voice composed of the same single syllables input by the voice input means into voices for each single syllable, and determining the single syllable intended by the speaker based on the respective voices Speech recognition device comprising a. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; And If the single-syllable voice input by the voice input means is any one of the sound of the sound, the sound of the sound, the sound of the sound, and the sound of the sound, the sound corresponding to the sound of sound, sound of the sound, sound of the sound, and sound of the sound of the sound is determined by the speaker. When the voice input by the voice input means is a specific word meaning a predetermined sound, the single syllable determined immediately before is changed to a single syllable of the corresponding sound sound, and the voice input by the voice input means is a specific word meaning a predetermined sound. In the case, the short syllable determined immediately before is changed to the corresponding syllable single syllable, and when the voice input by the voice input means is a specific word meaning a predetermined tactile sound, the short syllable determined immediately before is changed to the corresponding tactile single syllable, and the voice input is performed. If the voice input by the means is a specific word meaning a predetermined halftone sound, the single syllable determined immediately before Speech recognition means for changing a monosyllable corresponding bantak negative Speech recognition device comprising a. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; Storage means for storing a dictionary to which a single syllable corresponds to a combination of a single syllable recognition word consisting of a plurality of syllables; And Speech recognition means for analyzing the speech input by the speech input means to specify a combination of the specific syllables for single syllable recognition, and to determine the single syllable corresponding to the specific combination based on a dictionary stored by the storage means. Speech recognition device comprising a. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; Speech recognition means for specifying a candidate single syllable by analyzing the speech inputted by the speech input means; Display means having a display area for displaying designated information, and having a sensor for sensing a speaker's manipulation of the surface of the display area together with a position in the display area; And In the case where the speech recognition means specifies a plurality of candidate single syllables, each object corresponding to the candidate single syllables is displayed in the display area of the display means as the largest object group in the display area, and operated by the speaker to perform the Control means for determining the candidate single syllable corresponding to the object displayed at the position detected by the sensor as a single syllable intended by a speaker Speech recognition device comprising a. The method of claim 12, The control means selects and displays on the display means only three of each object displayed on the display means in high order in accordance with the prospect of candidate single syllables corresponding to the object. Speech recognition device. The method of claim 12, The control means for displaying each object displayed on the display means on the display means by changing a visual characteristic in accordance with the prospect of candidate single syllables corresponding to the object.  Speech recognition device.  The method of claim 12, The control means, at the time of the determination, handles a specific range that specifies each object by a position sensed by the sensor more than a display range occupied by the object in the display area of the display means.  Speech recognition device. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; Speech recognition means for specifying a candidate single syllable by analyzing the speech inputted by the speech input means; Display means having a display area for displaying designated information, and having a sensor for sensing a speaker's manipulation of the surface of the display area together with a position in the display area; And In the display area of the display means, an object representing a single syllable is displayed in correspondence with 50 notes, and when the speech recognition means specifies a plurality of the candidate single syllables, the object in the display area corresponding to each candidate single syllable is displayed. Single syllables intended for the candidate single syllables represented by the object displayed by changing the visual characteristics from other objects, and not being limited to objects whose visual characteristics have been changed but displayed by the speaker operated at the position detected by the sensor. Control means to determine as Speech recognition device comprising a. In a speech recognition apparatus mounted on a vehicle that determines a single syllable intended by a speaker based on a voice input by a speaker, Voice input means for inputting a voice spoken by the speaker; Speech recognition means for specifying a candidate single syllable by analyzing the speech inputted by the speech input means; Display means having a display area for displaying designated information and having a sensor for sensing a speaker's manipulation of the surface of the display area together with a position in the display area; Vehicle state acquiring means for acquiring information on whether the vehicle is traveling; The control means of the speech recognition apparatus according to claim 12; The control means of the speech recognition apparatus according to claim 16; And On the basis of the information acquired by the vehicle state acquiring means, when it is determined that the vehicle is running, the control means of the speech recognition apparatus according to claim 12 is functioned, and when it is determined that the vehicle is stopped, Main control means for functioning the control means of the described speech recognition apparatus Speech recognition device comprising a. In the speech recognition device for determining the single syllable intended by the speaker based on the voice input by the speaker, Voice input means for inputting a voice spoken by the speaker; Speech recognition means for specifying a candidate single syllable by analyzing the speech inputted by the speech input means; Notifying means for notifying specified information; Receiving means for receiving a speaker's operation; Storage means for storing a dictionary composed of a confirmation word composed of a plurality of syllables corresponding to each syllable; And The voice recognition means executes a notification process of notifying the notification means by the confirmation word stored by the storage means by the most likely candidate single syllable among the specific candidate single syllables, and performing an operation meaning a decision from the speaker. Control means for determining the candidate single syllable notified at the time of the notification process as the single syllable intended by the speaker when the receiving means accepts Speech recognition device comprising a. The method of claim 18, The storage means stores the plurality of dictionaries by dividing the dictionaries by types of the words for confirmation, The control means selects the dictionary stored by the storage means based on the operation of the speaker accepted by the reception means, and executes the notification process using the selected dictionary. Speech recognition device. The method of claim 18, The confirmation word in the dictionary stored by the storage means can be registered by a speaker in advance by operating the receiving means.  Speech recognition device. The method of claim 18, Display means having a display area for displaying specified information, and having a sensor for sensing a speaker's manipulation of the surface of the display area together with a position in the display area, The control means, when the speech recognition means specifies a plurality of candidate single syllables, displays the identification words corresponding to those candidate single syllables as objects on the display means as objects, and is operated by a speaker to be detected by the sensor. Determining the candidate single syllable corresponding to the object indicated at the position as a single syllable intended by the speaker. Speech recognition device. The method of claim 21, The control means, when the speech recognition means specifies a plurality of candidate single syllables, displays the identification words corresponding to the candidate single syllables as an object on the display means, and then displays the identification means specified by the speech recognition means. Determining the candidate single syllable corresponding to a word as a single syllable intended by a speaker  Speech recognition device. A program for causing a computer to function as at least one of speech recognition means or control means in the speech recognition apparatus according to any one of claims 1, 12 and 18. Computer-readable storage medium for storing data. A program for causing the computer to function as a speech recognition means in the speech recognition apparatus according to claim 6. Computer-readable storage medium for storing data. A program for causing the computer to function as a main control means in the speech recognition apparatus according to claim 17. Computer-readable storage medium for storing data. delete delete delete A navigation device for performing predetermined navigation processing, The voice recognition device according to any one of claims 1 to 22, A single syllable group intended by the speaker acquired by the speech recognition apparatus is used for the navigation processing. Navigation device.

Images