JPH11250078A - Guiding device having speech recognition function - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the performance of speech recognition and to improve convenience to use by reducing the number of vocalization for speech recognition. SOLUTION: A signal undergoes speech recognition processing in a speech processing part 2 and recognition results are given to a CPU 1 as a numeric value. The CPU 1 retrieves a word dictionary transition table 4 based on the recognition results, selects a dictionary to be used next from a word dictionary storing part 3 and sends it to the part 2. Also, it outputs guidance that urges vocalization to input a needed signal. Since a user inputs the next speech in accordance with a guidance output, the needed signal is inputted and it can easily be identified even if the same word exists plural times in a word dictionary.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information guide device with a voice recognition function for inputting a voice signal, recognizing a word contained in the voice signal, and guiding information stored in advance about the word.

[0002]

2. Description of the Related Art Heretofore, attempts have been made to perform input using audio signals during various types of information processing. When an audio signal is used, complicated operations such as using a keyboard are not required, and anyone can easily perform information processing.

[0003] In addition, the interface between the information processing apparatus and the person using the audio signal eliminates the need for the person to look at the display screen, unlike the combination of a keyboard and a display device. A person can use the information processing device such as a navigation device while driving a vehicle.

In a voice recognition function for recognizing a word contained in a voice signal, generally, a dictionary prepared in advance for a word is collated, and the word is specified according to the collation result. In the dictionary, for example, words are stored as combinations of phonemes. For combinations of phonemes extracted from the input speech signal, among combinations of phonemes stored in the dictionary,
A combination of phonemes with the highest similarity is selected, and a word corresponding to the combination of phonemes is recognized as a word included in the input speech signal.

[0005] In a navigation device, it is expected that an input of a destination or a waypoint, for example, is performed by a voice signal. Preferably, the entry of such a point can simply be made with the shortest words required. In a navigation device, a dictionary stores combinations (words) of phonemes corresponding to an enormous number of points so that guidance is possible even if the user moves over a wide range to some extent. If the length of the combination of phonemes to be compared becomes large (the number of words increases), the capacity required for the dictionary also increases, and it becomes impossible to store combinations of phonemes corresponding to many words. is there. However, when the target range is widened, for example, there may be a plurality of facilities and place names having the same name. In order to distinguish facilities having the same name, it is necessary to distinguish them by adding a prefecture name or the like in a higher hierarchy, or to separate a dictionary for each prefecture. In these cases, audio signals must be input in order from the upper hierarchy to the final facility name. What is necessary to have many dictionaries separately is a facility name, a telephone number, an address, or the like, which is a category having a large number of words.

[0006]

In the conventional speech recognition,
Since only words contained in the dictionary can be recognized, it is necessary to prepare all the words in which a human may utter a voice in one dictionary. As a result, the following problems occur in a category having an infinite number of words, such as a facility name.

[0007] The dictionary size per speech recognition dictionary increases. As the number of words in the dictionary for speech recognition increases, the time required for word matching increases. Due to the increase in the number of words in the dictionary for speech recognition, similar words increase, for example, "shiga" and "saga", and the recognition rate decreases. If dictionaries are divided for categories having a large number of words, there will be many dictionaries used for speech recognition, and the number of times the user of the device must utter will increase, resulting in discomfort.

It is an object of the present invention to reduce the size of a dictionary used for speech recognition, shorten the time required for word matching, improve the recognition rate, and reduce the number of times a user speaks. An object of the present invention is to provide a guidance device with a voice recognition function.

[0009]

SUMMARY OF THE INVENTION According to the present invention, information about a plurality of names is stored in advance for each name in one or more types, and information using information stored for a designated name designated by a user is provided. A voice recognition means for receiving a voice signal and recognizing a name included in the voice signal; and an information search for retrieving information stored in advance for a specified name recognized by the voice recognition means. Means, and when there are a plurality of types of information to be searched by the information search means, an output prompting voice input of an additional name for selecting one of the plurality of kinds is output, and the information is output in accordance with the additional name recognized by the voice recognition means Information selecting means for selecting only one type of information, information for which only one type is searched by the information searching means for the designated name input by the audio signal, or A guide device with speech recognition function which comprises an information guide means for guiding the information to be selected by one by distribution selection means.

According to the present invention, when a speech signal including a designated name is input to the speech recognition means, the information retrieval means retrieves information stored in advance for the designated name recognized by the speech recognition means. When there are a plurality of types of information searched by the information search means, the information selection means prompts the user to input an additional name by voice. When the additional name is recognized by the voice recognition means, only one type of information is selected from a plurality of types of information. Information searched for only one type by the information search means or information selected only by the information selection means is guided by the information guide means. The user first speaks about the designated name of the object, and can immediately receive guidance when there is only one type of corresponding information. When a plurality of types of information are stored, an utterance for inputting additional information prompted by the information selecting means may be performed. Therefore, information can be comfortably received while minimizing the number of utterances. Since the dictionary used for information recognition may store a plurality of types of information for the same designated name, shorten the combination of phonemes (reduce the number of words) to reduce the dictionary size. In addition, the time required for word matching can be reduced. When a plurality of types of information are stored for the same designated name, an additional name is input and selected, so that the final word recognition rate can be improved.

Further, the present invention is an apparatus for preliminarily storing one or more types of information for a plurality of names for each name, and providing guidance using information stored for a designated name designated by a user. A voice recognition means for receiving a voice signal and recognizing a name included in the voice signal; and inputting an additional name for limiting the specified name when inputting the specified name with the voice signal to the voice recognition means. A voice processing unit that prompts the user, an information search unit that searches for information stored in advance for a designated name recognized by the voice recognition unit, and a plurality of types of information that are searched by the information search unit. An information selection means for selecting only one type of information according to the recognized additional name; A guide device with speech recognition function which comprises an information guide means for guiding the information to be selected by one the information is retrieved by the class or information selecting means.

According to the present invention, when a voice signal is input to the voice recognition means to instruct retrieval and guidance of information based on the specified name, the voice processing means uses the voice signal of an additional name for limiting the specified name. Prompting the user to input a required additional name. Even if a plurality of types of information are stored for the same designated name, only one type can be selected because the additional name is input. Since a dictionary used for voice recognition may correspond to a plurality of types of information for the same designated name, the combination of phonemes can be shortened, the dictionary size can be reduced, and the word matching time can be reduced. Can also be reduced. One type of information can be selected from a plurality of types of information by the additional name, so that the final speech recognition rate is improved,
Since only the necessary additional name needs to be input, the number of utterances can be reduced to make the device easier to use.

Still further, the present invention is an apparatus for performing guidance using information stored for a designated name specified by a user, in which one or more types of information for a plurality of names are stored in advance for each name. A voice signal is input, voice recognition means for recognizing a name included in the voice signal, temporary storage means for temporarily storing a voice signal for the specified name, and a specified name following the specified name. When a voice signal for the additional name to be limited is input to the voice recognition means, the voice processing means for recognizing the voice signal for the specified name stored in the temporary storage means after the additional name is recognized by the voice recognition means. For a specified name recognized by the voice recognizing means, the pre-stored information is searched by the searching means, and the information searched by the information searching means is divided into a plurality of types. Information selecting means for selecting only one type of information according to the additional name recognized by the voice recognizing means, and information which is searched for only one type by the information searching means for the specified name input by the voice signal, or An information guiding means for guiding only one type of information selected by the selecting means.

According to the present invention, the voice processing means temporarily stores the previously input designated name in the temporary storage means, and after the additional name is input and recognized, stores the specified name in the temporary storage means. Since the stored specified name is recognized by the voice recognition means, even when a plurality of dictionaries are used, or when a plurality of types of information correspond to the same specified name in one dictionary, the dictionary selection or the By selecting the type of information, information corresponding to the word specified by the additional name and the designated name can be easily guided. Even if a plurality of types of information for the same designated name are included in the same dictionary or another dictionary, the information can be easily selected by additional information. And the number of words is reduced, so that the dictionary size can be reduced. Since dictionary selection or word collation is performed based on the additional name and the designated name, the time required for word collation can be reduced, and the recognition rate can be improved. Since the user only has to say the designated name first and then say the additional name later, the user can comfortably input the voice.

[0015] The present invention further comprises dictionary means for preliminarily storing a voice signal as a reference for the word corresponding to the designated name and the additional name, and the voice recognition means includes a word and a dictionary included in the voice signal. Compare with a reference audio signal to be stored, output a numerical value indicating the degree of mismatch as a comparison result, the temporary storage means, when the numerical value representing the comparison result from the voice recognition means exceeds a predetermined reference value And temporarily storing the audio signal.

According to the present invention, the dictionary means previously stores a voice signal as a reference for the designated name and the additional name, and the voice recognition means compares the words contained in the voice signal with the dictionary means. , A numerical value representing the degree of mismatch is stored as the comparison result. The temporary storage means temporarily stores the voice signal when the numerical value representing the collation result from the voice recognition means exceeds a predetermined reference value. In addition, the possibility of erroneous recognition can be reduced, and the recognition rate can be improved by inputting an additional name. Once entered, the name is temporarily stored, so there is no need to enter it again, and it can be used effectively for collation of dictionary means.

Also, in the present invention, the dictionary means has a plurality of types of dictionaries corresponding to the designated name or the additional name, and the voice recognizing means comprises a dictionary for selecting the word and one dictionary. According to the present invention, a dictionary to be used next is selected in accordance with a numerical value representing a collation result.

According to the present invention, the dictionary means is provided with a plurality of types of dictionaries corresponding to the designated name or the additional name, and the speech recognition means is provided with a numerical value representing the result of matching the word with the dictionary means having selected one dictionary. , A dictionary to be used next is selected, so that an appropriate dictionary can be easily selected from a plurality of dictionaries.

Further, in the present invention, the voice recognizing means performs a temporary storage of the voice signal including the word by the temporary storing means and a selection of a dictionary by the dictionary means in accordance with a numerical value representing a result of collation of the word. And comparing the words contained in the temporarily stored voice signal using the selected dictionary.

According to the present invention, the voice recognition means temporarily stores a voice signal including a word and selects a dictionary by the dictionary means in accordance with a numerical value representing the result of word matching. The selected dictionary is collated using the words contained in the temporarily stored audio signal, so there is no need to input the same word again, and speech recognition using multiple dictionaries using the once input word is performed. It can be done efficiently.

Further, in the present invention, the voice recognition means checks a word contained in the temporarily stored voice signal using a selected dictionary after the voice signal is input a predetermined number of times. It is characterized by performing.

According to the present invention, the temporarily stored voice signal is compared with the dictionary after a plurality of voice inputs.
There is no need to input a new voice signal after a plurality of voice inputs, and a word input in a plurality of times can be used as an additional name for reliable voice recognition.

According to the present invention, there is provided a result storage means for storing, in chronological order, the suitability of the recognition result of the voice recognition means with respect to the designated name, Updating means for updating each of the values so as to decrease or increase the value.

According to the present invention, the suitability of the recognition result of the voice recognition means for the designated name is stored in the result storage means in time series, and the updating means is used as a reference based on the suitability of the stored recognition result. Since the value is updated so as to decrease or increase, the reference value is lowered when the hit rate is high and there is no misrecognition, and the reference value is increased when the hit rate is low, and accurate and reliable recognition results are obtained. It can be adjusted to obtain.

Further, the present invention is characterized in that it further comprises setting means for setting the reference value.

According to the present invention, since the reference value can be set by the setting means, the user can set a more appropriate reference value based on the hit rate of the result of voice recognition.

In the present invention, the designated name is a facility name, the additional name is a prefecture name, and the information guide means guides map information.

According to the present invention, when providing map information, a facility name is input as a designated name by voice, and a prefecture name is added as an additional name, so that the same facility name exists in a plurality of prefectures. However, it can be easily distinguished.

[0029]

FIG. 1 shows a schematic electrical configuration of a navigation device with a voice recognition function as a first embodiment of the present invention. The CPU 1 controls the entire operation of the navigation device with a voice function according to a predetermined program. The voice recognition process is performed by the voice processing unit 2 using the word dictionary stored in the word dictionary storage unit 3. The word dictionary used for the speech recognition processing is selected according to the word dictionary transition table 4. The input of the audio signal to the audio processing unit 2 is performed by analog / digital conversion (hereinafter, “A”).
/ D) is supplied to the circuit 5 with an audio signal from the microphone 6 and converted into digital data. The voice processing unit 2 has a voice synthesizing function in order to provide voice guidance to the user of the present device to prompt the user to speak. The digital data subjected to voice synthesis from the voice processing unit 2 is converted into an analog voice signal by a digital / analog conversion (hereinafter, abbreviated as “D / A”) circuit 7, and sound output from a speaker 8.

The CPU 1 serves as a navigation device, a CD / ROM in which map information such as road data is stored in advance.
With reference to a recording medium 9 such as a DVD or a DVD, related map information can be read in accordance with words such as place names recognized by the audio processing unit 2. Further, the position detecting section 10 detects a current position of a vehicle or the like, and the CPU 1 can also read map information around the current position recorded on the recording medium 9 based on the detected current position. When the navigation device with a voice function is mounted on a vehicle, the position detection unit 10 estimates the current position based on an estimation method based on signals from a speed sensor, a direction sensor, or the like;
A detection result of the current position using the PS can be used. The CPU 1 stores the current position detected by the position detection unit 10, map information read from the recording medium 9, and the like.
The image is displayed on the display unit 11 as an image. In addition, as a navigation device, a route from a departure point to a destination can be searched or set in advance, and while being mounted on a vehicle and moving, an image using the display unit 11 or a speaker 8 can be used. Route guidance can also be provided. In the navigation device with a voice recognition function according to the present embodiment, for example, a voice signal can be used for inputting a destination, a departure place, or a waypoint on the way that must be input when setting a route.

FIG. 2 shows a voice recognition processing procedure of the navigation apparatus with a voice recognition function of FIG. When an audio signal is input to the microphone 6 in step a1, the audio processing unit 2
Then, a word included in an input speech signal is recognized based on a word dictionary provided in advance by the CPU 1 or provided as a default. In step a2,
The CPU 1 receives the recognition word result from the voice processing unit 2.
At this time, a numerical value representing the degree of coincidence is also received. In step a3, the CPU 1 searches the word dictionary transition table 4, and determines in step a4 whether there is a next selection dictionary which is a word dictionary to be selected next. For example, if the word dictionary used in the voice processing unit is a word dictionary for one facility name and the same facility name exists in other prefectures, the result of the word dictionary transition table search determines that there is a next selection dictionary. Is done. At this time, the process proceeds to step a5,
The CPU 1 stores the next word dictionary to be selected in the word dictionary storage unit 3.
And sends it to the voice processing unit 2.

Further, at step a6, a guidance output for prompting the next utterance is made via the voice processing unit 2. After the output in step a6 is performed, or when it is determined in step a4 that there is no next selected dictionary, 1 is output in step a7.
The process of recognizing the words included in the one voice signal is ended. When the user responds to the guidance output in step a6 and voice input is performed by the next utterance, the processing from step a1 is performed again. It should be noted that a flag indicating the presence / absence of a dictionary to be selected next may be provided in the word dictionary and used instead of the word dictionary transition table.

FIG. 3 shows a schematic electrical configuration of a navigation device with a voice recognition function according to a second embodiment of the present invention. In the present embodiment, portions corresponding to the embodiment of FIG. 1 are denoted by the same reference numerals, and redundant description will be omitted. In the present embodiment, the audio processing unit 2 includes a temporary storage memory 13 and a counter 14 realized by a memory 12 such as an internal RAM and a buffer. A preset number N is set in the counter 14, and when the number of voice inputs reaches N, the phonetic phonemes included in the input voice signal are temporarily stored in the temporary storage memory 13. Further, the number of times and the like can be input from the input unit 15 such as a keyboard.

FIG. 4 shows a speech recognition procedure of the embodiment of FIG. When the Nth utterance indicated by the counter 14 is input, the recognition process starts from step b1. In step b2, the voice processing unit 2 temporarily stores the uttered phoneme of the uttered word in the temporary storage memory 13. In step b3,
The CPU 1 refers to the dictionary transition table 4 and selects the next selected dictionary and transmits it to the voice processing unit 2. In step b4, the process waits for the input of the (N + 1) -th uttered phoneme. When the input phoneme is input, the process refers to the dictionary transition table 4 in step b5 to select the next selection dictionary corresponding to the word recognized in the (N + 1) -th Transmission to the audio processing unit 2 is performed. In step b6, the voice processing unit 2 calls the uttered phoneme temporarily stored in the temporary storage memory 13. In step b7, the speech processing unit 2 compares the called uttered phoneme with the word dictionary in the word matching unit, outputs a numerical value indicating the degree of disagreement with the recognition result as a result, and ends the recognition processing in step b8. In the processing of the present embodiment, it is necessary to determine in advance to which utterance word this method is to be applied and the value of N in the counter 14 in advance by determining the value in the program of the CPU 1.

FIG. 5 shows a third embodiment of the present invention.
For example, in the navigation device with a voice recognition function having the same electrical configuration as that of FIG. 3, a processing procedure for determining whether to use the recognized word of the verification result based on the word verification result in the voice processing unit 2 will be described. When a voice signal is input in step c1, the CPU 1 receives the result of collating the uttered word from the voice processing unit 2 in step c2. At this time, a numerical value indicating the degree of matching between the uttered word and the word in the word dictionary is also received. In step c3, it is determined whether or not the numerical value received as the collation result is larger than a preset value. This numerical value represents the number of elements that do not match, and the higher the numerical value, the lower the degree of matching. When it is determined that the comparison result is not larger than the set value, a result transmission required signal is transmitted to the audio processing unit 2 in step c4. When it is determined that the comparison result is larger than the set value, a transmission result unnecessary signal is transmitted to the audio processing unit 2 in step c5. After step c4 or step c5, the process of determination is ended in step c6. By such a determination process, it is possible to prevent guidance based on a word having a low degree of matching or selection of a selection dictionary.

FIG. 6 shows a fourth embodiment of the present invention.
When the numerical value indicating the result of performing the word matching on the uttered word is larger than the set value in the navigation device with the voice recognition function having the electrical configuration as shown in FIG. 3, another word dictionary is automatically selected. When a voice signal is input in step d1, the CPU 1 receives a result of collating the uttered word from the voice processing unit 2 in step d2. In step d5, it is determined whether or not the numerical value accompanying the processing result is larger than the set value. When it is determined that it is not large, the dictionary transition table 4 is searched based on the word recognized in step d5, A is selected as the next selected dictionary in step d5, and the dictionary A is transmitted to the voice processing unit 2. When it is determined in step d3 that the comparison result is larger than the set value, the dictionary transition table 4 is searched in step d6, and the next selection dictionary different from the next selection dictionary A selected in step d5 as the next selection dictionary in step d7. B is selected and transmitted to the audio processing unit 2. When the processing in step d5 or step d7 ends, the selection processing of the next selection dictionary ends in step d8.

FIG. 7 shows a fifth embodiment of the present invention.
In the navigation device with a voice recognition function similar to that of FIG. 3, only when the numerical value representing the result of performing the word matching on the uttered word in the voice processing unit 2 is larger than the set value, the uttered phoneme is temporarily stored. Automatically selects the dictionary to use for
The procedure for performing the next word matching using the uttered phoneme will be described. When a voice is input by utterance in step e1, the CPU 1 receives a result of the utterance word collation from the voice processing unit 2 in step e2. In step e3, it is determined whether or not the numerical value accompanying the collation result is larger than the set value. When it is determined that it is not large, in step e4, the sound processing unit 2
A result transmission necessary signal is transmitted to the dictionary transition table 4 in step e5. At step e6, the next selection dictionary A is selected and the dictionary A is transmitted to the voice processing unit 2.

When it is determined in step e3 that the comparison result is larger than the set value, a transmission result unnecessary signal is transmitted to the voice processing unit 2 in step e7. In step e8, a dictionary transition table is searched, and in step e9, step e
The selection of the next selection dictionary B different from the next selection dictionary A selected in step 6 and the transmission of the dictionary B to the voice processing unit 2 are performed.

When the processing in step e9 is completed, the phonetic phoneme temporarily stored in the temporary storage memory 12 in step e10 is called. The phonetic phoneme called in step e11 is compared with the word dictionary by the word matching unit of the voice processing unit 2,
The procedure ends at step e12. After step e6, the procedure ends at step e12.

FIG. 8 shows a sixth embodiment of the present invention.
In the navigation device with a voice recognition function as shown in FIG. 3, after inputting a preset M-th voice signal, a further J-time voice signal is input, and then a word is stored using the temporarily stored voice signal. The procedure for performing the matching will be described. Steps f1 to f9 are substantially the same as steps e1 to e9 in the fifth embodiment of FIG. In step f10, the CP
U1 determines whether or not the M + J-th recognition word result has been received from the voice processing unit 2. If it is determined that the dictionary has not been received, the next selected dictionary is selected from the word dictionary transition table 4 and transmitted to the voice processing unit 2, and the process returns to step f10. If it is determined in step f10 that the M + J-th recognition word result has been received, step f1
In step 2, the next selected dictionary is selected from the dictionary transition table corresponding to the recognized word and transmitted to the voice processing unit 2. The voice processing unit 2 calls the uttered phoneme temporarily stored in the temporary storage memory 12 in step f13, matches the uttered phoneme called in step f14 with the word dictionary in the word matching unit, and ends the matching procedure in step f16. I do. The values of M and N are input, for example, from the input unit 15 by the user.

FIG. 9 shows a seventh embodiment of the present invention.
A procedure for updating a threshold used for voice recognition in the voice processing unit 2 using the navigation device with a voice recognition function as shown in FIG. 3 will be described. When a voice signal is input in step g1, the uttered phoneme called in step g2 is compared with the word dictionary. In step g3, it is determined whether or not the user has corrected the recognition result for the recognized word. If it is determined that the correction has been made, it is determined in step g4 whether the correction is the first correction. If it is determined that this is the first correction, at step g5 the word correction threshold value corrected at that time is stored in the memory.
When it is determined in step g4 that the word is not the first correction, in step g6, the threshold values of the words corrected up to that time are averaged and stored. When step g5 or step g6 is completed, the process proceeds to step g7, and it is determined whether or not a preset period has elapsed. If it is determined that the period has elapsed, the process proceeds to step g8, and the threshold used for the comparison is rewritten to the threshold stored in the memory. Step g
When it is determined in step 3 that there is no word correction, or when it is determined in step g7 that the set period has not elapsed,
Alternatively, after rewriting is performed in step g8, the procedure ends in step g9.

FIG. 10 shows, as an eighth embodiment of the present invention, a partial electrical configuration of a navigation device capable of setting a threshold value of a voice recognition process by a user. Input device 1
Reference numeral 5 denotes a remote control device usually used for a navigation device, which includes a joystick, a numeric keypad, and the like.
In an in-vehicle navigation device, the input device 15 such as a remote controller has only limited functions.
The PU 1 sets the threshold value in the threshold value setting unit 16 while displaying the setting screen on the display unit 11.

FIG. 11 shows a procedure for changing the threshold value in the navigation device shown in FIG. Steps h1 to h8 are the same as steps g1 to g8 shown in FIG. In the present embodiment, when it is not determined in step h7 that the set period has elapsed, it is determined in step h9 whether a manual change using the input device 15 has been performed by the user. If it is determined that the operation has been performed, step h1
At 0, the contents of the threshold value stored in the memory are cleared, and the manually changed threshold value is used as it is. Step h3
When it is determined in step h9 that there is no correction, when it is determined that there is no change in step h9, or when the procedure of step h8 or step h10 ends, step h1
In step 1, the processing procedure ends.

In each of the embodiments described above, the case where the present invention is applied to a navigation device with a voice recognition function is described. However, the present invention can be similarly applied to a device for guiding various information in accordance with voice input. In particular, when a name serving as a key for information retrieval is specified in multiple hierarchies, voice input can be performed with a minimum necessary utterance to efficiently receive information guidance.

[0045]

As described above, according to the present invention, only relevant information is searched for by utterance of only a designated name, and when a plurality of types of information are stored in advance, utterance of an additional name is performed. Since only one type of information is selected, the number of utterances can be reduced and the usability of the device can be improved. The length of the combination of phonemes of the word to be recognized as the specified name is shortened to reduce the number of words, and even if multiple types of information are stored in the same word, it is possible to select with an additional name. Thus, the word matching time can be reduced, and the recognition rate can be improved.

Further, according to the present invention, when the designated name is input by a voice signal, the input of the additional name for limiting the specified name by the voice signal is prompted. The corresponding information can be searched and selected, only necessary words can be input, and the usability can be improved by reducing the number of utterances. In addition, by shortening the length of phonemes for collating words, the dictionary size and collation time can be reduced, and the recognition rate can be improved.

Further, according to the present invention, if the designated name is first spoken, and the additional name is spoken and voice input is performed, the speech recognition is performed first for the additional name, and the voice is spoken for the temporarily stored designated name. Since the information is recognized and guidance of the corresponding information is performed, it is not necessary to input the designated name again after the input of the additional name, and the number of utterances can be reduced and the usability can be improved. Further, even if a plurality of types of information are stored for the same word, one type can be selected from a plurality of types by an additional name, so that the dictionary size and word matching time can be reduced, and the recognition rate can be improved. Can be planned.

Further, according to the present invention, the words are temporarily stored when the result of collation of the words contained in the voice signal with the dictionary means exceeds a predetermined reference value. Therefore, since the temporarily stored audio signal is used, it is not necessary to repeatedly input the audio signal, and the recognition rate and the usability can be improved.

Further, according to the present invention, a plurality of types of dictionaries can be appropriately selected, so that the size per dictionary can be reduced, the collation time can be reduced, and the word recognition rate can be reduced. Improvement can also be achieved.

According to the present invention, temporary storage of the voice signal and selection of a dictionary for matching the words are performed in accordance with the numerical value representing the verification result, and the words included in the voice signal temporarily stored for the selected dictionary are determined. Since the matching is performed using the word, there is no need to repeat the input of the word, and the recognition rate of the matching can be improved by using an appropriate dictionary among the plurality of dictionaries.

Further, according to the present invention, the audio signal temporarily stored based on the result of collation with the dictionary is collated again with the dictionary after inputting a predetermined number of audio signals. The recognition rate can be improved by using the input result of the performed audio signal.

Further, according to the present invention, the reference value for determining whether to temporarily store the voice signal is updated based on the time-series storage of the appropriateness of the voice recognition result. Can be automatically adjusted to

Further, according to the present invention, the reference value for judging whether or not to temporarily store the input signal by voice can be set from the setting means, so that the user of the apparatus can recognize the signal more appropriately. A reference value can be set to obtain a result.

Further, according to the present invention, map information is guided by facility names in an easy-to-understand manner. When the same facility name exists in different prefectures, etc., each prefecture name is input as an additional name to ensure the purpose. Guide of map information about the facility name to be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic electrical configuration of a navigation device with a voice recognition function according to a first embodiment of the present invention.

FIG. 2 is a flowchart illustrating a voice recognition processing procedure according to the embodiment of FIG. 1;

FIG. 3 is a block diagram showing a schematic electrical configuration of a navigation device with a voice recognition function according to a second embodiment of the present invention.

FIG. 4 is a flowchart illustrating a voice recognition processing procedure according to the embodiment of FIG. 3;

FIG. 5 is a flowchart showing a voice recognition processing procedure according to a third embodiment of the present invention.

FIG. 6 is a flowchart illustrating a voice recognition processing procedure according to a fourth embodiment of the present invention.

FIG. 7 is a flowchart illustrating a voice recognition processing procedure according to a fifth embodiment of the present invention.

FIG. 8 is a flowchart illustrating a voice recognition processing procedure according to a sixth embodiment of the present invention.

FIG. 9 is a flowchart showing a procedure for automatically correcting a threshold value in speech recognition according to a seventh embodiment of the present invention.

FIG. 10 is a block diagram showing a schematic electrical configuration of a navigation device with a voice recognition function according to an eighth embodiment of the present invention.

11 is a flowchart showing a processing procedure of the embodiment of FIG.

[Explanation of symbols]

 Reference Signs List 1 CPU 2 Voice processing unit 3 Word dictionary storage unit 4 Word dictionary transition table 6 Microphone 8 Speaker 11 Display unit 12 Memory 13 Temporary storage memory 14 Counter 15 Input device 16 Threshold setting unit

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FIG06F 15/403 350C

Claims (10)

[Claims] 1. A device for performing guidance using information stored about a designated name designated by a user, in which one or more kinds of information about a plurality of names are stored in advance for each name. A signal is input, and a voice recognition unit that recognizes a name included in the voice signal, an information search unit that searches information stored in advance for a specified name that is recognized by the voice recognition unit, and a search that is performed by the information search unit When there are a plurality of types of information to be output, an output for prompting a voice input of an additional name for selecting one of the plurality of types is performed, and information selection for selecting only one type of information according to the additional name recognized by the voice recognition unit. Means, and information to be searched for only one type by the information searching means, or one kind of information to be specified by the information selecting means, for the designated name inputted by the audio signal. And a data guiding means for guiding information selected only by a class. 2. A device for preliminarily storing one or more types of information for a plurality of names for each name, and providing guidance using information stored for a specified name specified by a user. A voice recognition means for receiving a signal and recognizing a name included in the voice signal, and a voice for prompting the input by a voice signal of an additional name for limiting the specified name when the specified name is input to the voice recognition means by the voice signal A processing unit; an information search unit for searching information stored in advance for a designated name recognized by the voice recognition unit; and a plurality of types of information searched by the information search unit. An information selection unit that selects only one type of information according to the additional name, and only one type of information is searched by the information search unit for a specified name input by a voice signal. The information or information selection means by the voice recognition function with guide apparatus which comprises an information guide means for guiding the information to be selected only one type. 3. A device for preliminarily storing one or more types of information for a plurality of names for each name, and providing guidance using information stored for a specified name specified by a user. A voice recognition means for receiving a signal and recognizing a name included in the voice signal, a temporary storage means for temporarily storing a voice signal for the specified name, and an additional name for limiting the specified name following the specified name Voice processing means for recognizing, by the voice recognition means, the voice signal for the designated name stored in the temporary storage means after the additional name is recognized when the voice signal is input to the voice recognition means; When there are a plurality of types of information to be retrieved by the information retrieval means and for the designated name recognized by the Information selecting means for selecting only one type of information according to the additional name recognized by the means; information for searching only one type by the information searching means, or one type of information for the specified name inputted by the voice signal; An information guide means for guiding information selected only by the user. 4. A dictionary unit for storing in advance a reference voice signal for a word corresponding to the designated name and the additional name, wherein the voice recognition unit stores a word included in the voice signal and the dictionary unit. Comparing with a reference audio signal, outputting a numerical value indicating the degree of mismatch as a collation result, the temporary storage means, when the numerical value representing the collation result from the voice recognition means exceeds a predetermined reference value, the voice The guidance device with a voice recognition function according to claim 3, wherein the signal is temporarily stored. 5. The dictionary means has a plurality of types of dictionaries corresponding to the designated name or the additional name, and the voice recognizing means determines a matching result between the word and a dictionary means that has selected one dictionary. 5. The guidance device with a voice recognition function according to claim 4, wherein a dictionary to be used next is selected according to the numerical value represented. 6. The voice recognizing means performs temporary storage of the voice signal including the word by the temporary storage means and selection of a dictionary by the dictionary means in accordance with a numerical value representing the collation result of the word. 6. The guidance device with a voice recognition function according to claim 5, wherein the matching is performed on the words included in the stored voice signal using the selected dictionary. 7. The voice recognition means performs a collation using a selected dictionary with respect to a word included in the temporarily stored voice signal after the voice signal is input a predetermined number of times. The guidance device with a voice recognition function according to claim 6. 8. A result storage means for storing, in chronological order, the suitability of the recognition result of the speech recognition means for the designated name, and the reference value is set based on the suitability of the recognition result stored in the result storage means. The guidance device with a voice recognition function according to any one of claims 4 to 7, further comprising updating means for updating each of the information so as to decrease or increase the number. 9. The guidance device with a voice recognition function according to claim 4, further comprising a setting unit for setting the reference value. 10. The apparatus according to claim 1, wherein the designated name is a facility name, the additional name is a prefecture name, and the information guiding means guides map information. A guidance device with a voice recognition function according to item 1.