JP2004061754A - Voice controlled unit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a voice controlled unit which makes it easy to erase a part inputted by mistake when an address is vocally inputted. <P>SOLUTION: When a user speaks "return" successively twice, a display control part of the voice controlled unit erases a display of names of addresses on the display screen of a display part at a time. Consequently, when vocal input is redone from a prefecture name after, for example, "Aichi-prefecture", "Kariya-city", "Showa-machi", and "1-chome" are inputted, the user speaks "return" successively twice to erase a display of a plurality of words of "Aichi-prefecture, Kariya-city, Showa-cho, 1-chome" at a time. Consequently, the user can easily erase the display of the address without repeatedly voicing "return" again and again. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a voice control device.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, there is a voice control device that recognizes a voice uttered by a speaker and inputs an address. In this voice control device, when a user inputs a desired address, the desired address is delimited for each layer in units of names such as prefectures, municipalities, towns and villages (hereinafter referred to as layers). Read it out. For example, when inputting an address “1-1, Showa-cho, Kariya-shi, Aichi”, the first level “Aichi prefecture”, the second level “Kariya”, and the third level “Showa” The text is read out for each level, such as "town", "1st-chome" as the fourth level, and "1st address" as the fifth level.
[0003]
On the other hand, in the case where the layer portion that has been erroneously input is corrected and the input is performed again, the user issues “return” corresponding to the command for deleting the display of one layer, thereby deleting only one layer. For example, as shown in FIG. 15, when the user utters up to “Showa-cho 1-chome, Kariya-shi, Aichi-ken” and then inputs again from “Kariya-shi Aichi-ken”, the user must input “1-chome”. Immediately after the utterance, "return" is uttered twice in succession, whereby the address input to the target hierarchy is deleted. Thereafter, the user again inputs the address after “Kariya City, Aichi Prefecture”.
[0004]
As described above, in the conventional voice control device, the user reads the address divided into layers and inputs the desired address, and when erasing the erroneously input layer portion, returns “return”. Speak to erase the error.
[0005]
[Problems to be solved by the invention]
However, the command corresponding to “return” in the above-described voice control device is limited to one layer to be deleted. Therefore, for example, after inputting “1-chome, Showa-cho, Kariya-shi, Aichi”, when re-entering the voice from the prefecture name which is the first hierarchy, the user utters “return” four times consecutively. It is troublesome for the user to continuously utter such identical utterance contents.
[0006]
The present invention has been made in view of such a problem, and an object of the present invention is to provide a voice control device capable of easily deleting an address input by voice.
[0007]
[Means for Solving the Problems]
The voice control device according to claim 1, wherein voice input means for inputting voice, voice recognition means for recognizing a user's uttered voice input to the voice input means, and when voice is input by the user, Recognized utterance contents are classified and stored in a plurality of layers, and when a correction command for correcting the utterance contents is input, a part of the utterance contents classified and stored in the lowest layer is erased. Control means for deleting all of the utterance contents classified and stored in a plurality of layers when the correction command is input twice consecutively.
[0008]
As described above, when the correction command is input twice consecutively, the voice control device of the present invention deletes all of the utterance contents classified and stored in a plurality of layers. This makes it possible to easily delete the utterance content without repeatedly issuing the correction command.
[0009]
As described in claim 2, the utterance content classified and stored in a plurality of layers may be at least an address. For example, when the address is input by voice, after inputting four levels of “Aichi prefecture”, “Kariya city”, “Showa-cho”, and “1-chome”, the prefectural level of the first level is again input. In the case of redoing the voice input from the user, the user utters the command for correcting the address input by voice twice in succession, and the command is stored in four levels of “1-chome, Showa-cho, Kariya-shi, Aichi”. It is possible to delete all addresses. As a result, the user can easily delete all the input addresses without repeatedly issuing the correction command.
[0010]
According to the third aspect of the present invention, the addresses are classified into at least five levels: prefectures, municipalities, towns and villages, streets and addresses. Accordingly, when inputting the address by voice, the user can input the desired address by speaking the respective layers of the prefecture, city, county, town, village, character, and street of the desired address. Can be.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a voice control device according to an embodiment of the present invention will be described with reference to the drawings. In this embodiment, an example in which the voice control device of the present invention is applied to a car navigation device will be described.
[0012]
FIG. 1 is a block diagram illustrating a schematic configuration of a car navigation device according to the present embodiment. As shown in FIG. 1, the car navigation device 1 according to the present embodiment includes a voice recognition unit 10, a route guidance unit 11, and a vehicle position / vehicle direction calculation unit 12. The car navigation device 1 has a road map drawing unit (not shown) and the like. Further, the car navigation device 1 is connected to a microphone 2, a talk switch 3, a display device 4, a speaker 5, a GPS receiver 6, a vehicle speed sensor 7, a yaw rate sensor 8, a map database 9, and the like used for voice input. .
[0013]
The microphone 2 and the talk switch 3 are devices used for voice input. When inputting voice, for example, by pressing a push button of the talk switch 3, an input trigger signal is transmitted to a voice recognition unit 10 described later. The mode is changed from 2 to a mode for receiving a voice input.
[0014]
In the mode in which the voice input is accepted, when a voice is input by the user, the voice is converted into a voice signal by the microphone 2 and sent to the voice recognition unit 10. The voice recognition unit 10 recognizes the voice signal, converts the voice signal into a name corresponding to the voice, such as a prefecture or a city, and gives the name to the route guidance unit 11. For example, a voice recognized as “Aichiken” is converted into a prefecture name “Aichi”. The route guidance unit 11 receiving the name of the prefecture stores the received name and displays the name on the display device 4. Thereafter, each time the name such as a city, a ward, a town character, a street address, and an address is received, the route guidance unit 11 receives these names. Are hierarchically connected and stored, and displayed on the display device 4. In the present embodiment, prefectures are referred to as a first hierarchy, municipalities are referred to as a second hierarchy, towns and villages are referred to as a third hierarchy, streets are referred to as a fourth hierarchy, and addresses are referred to as a fifth hierarchy.
[0015]
Further, when the route guidance unit 11 receives all addresses including names such as prefectures, municipalities, towns and villages, streets and addresses, it searches for a point on a road map corresponding to this address, and searches for it. A mark indicating a point is displayed on the display device 4 together with a road map around the point.
[0016]
Further, when the user utters, for example, "return" during input of the address, the voice is recognized, converted into a command code corresponding to the voice, and given to the route guidance unit 11 and the like. For example, a voice recognized as “return” is converted into a command code of “display erase”. Upon receiving this command code, the route guidance unit 11 determines whether or not this command code has been received twice consecutively. Of the addresses to be deleted, only the name located at the lowest level is deleted. If the address is received twice consecutively, all of the stored and displayed addresses are collectively erased.
[0017]
The display device 4 is configured by a liquid crystal display that displays a road map and the like. Further, a touch panel may be employed as the display of the display device 4.
[0018]
The speaker 5 is used for outputting voice guidance, various warning sounds, and the like. For example, the speaker 5 may be a speaker provided in a vehicle or may be a built-in speaker in the car navigation device 1. .
[0019]
As is well known, the GPS receiver 6, the vehicle speed sensor 7, and the yaw rate sensor 8 generate a signal (hereinafter, referred to as a sensor signal) necessary for calculating the current position of the vehicle, the traveling direction of the vehicle, and the like. . The generated sensor signal is sent to the vehicle position / vehicle direction calculator 12.
[0020]
The map database 9 is stored in a storage medium (not shown), and includes map information and road information. As a storage medium, a CD-ROM or a DVD-ROM is generally used because of its data amount, but a medium such as a memory card or a hard disk may be used. The map information is data necessary for drawing landmarks and the like to be displayed on the display device 4, and is composed of data in which facility names, addresses, telephone numbers, coordinates on a map, and the like are associated.
[0021]
Next, the speech recognition unit 10 built in the car navigation device 1 will be described with reference to FIG. As shown in FIG. 1, the speech recognition unit 10 includes an AD conversion circuit 101, a recognition program processing unit 102, an acoustic model storage unit 103, a recognition dictionary storage unit 104, and the like.
[0022]
The AD conversion circuit 101 receives an analog audio signal input via the microphone 2, and converts this signal into a digitized signal. The converted digital audio signal is transmitted to the recognition program processing unit 102.
[0023]
The recognition program processing unit 102 uses the acoustic model storage unit 103 and the recognition dictionary storage unit 104 to convert a digital voice signal into a name of a prefecture or a command code. First, the recognition program processing unit 102 uses the technique such as the well-known hidden Markov model (Hidden Markov Model) stored in the acoustic model storage unit 103 to generate the utterance content (hereinafter, referred to as the digital speech signal 106). (Referred to as recognition word reading).
[0024]
The analyzed recognition word reading is compared with the recognition word stored in the recognition dictionary storage unit 104, and the most probable recognition word and the name of the prefecture, city, county or the like corresponding to the recognition word, or the command code Is extracted.
[0025]
Here, the recognition dictionary storage unit 104 will be described. The recognition dictionary storage unit 104 includes a prefectural dictionary that stores names of prefectures and recognition words in association with each other, a city dictionary that stores names of cities and districts divided by prefecture and recognition words in association with each other, A town and village character dictionary that is divided by prefecture and divided by city and district in a specific prefecture and stores the name and recognition word in association with each other, and the names of streets and streets are associated with recognition words It has a total of five recognition dictionaries, such as a chome dictionary and an address dictionary.
[0026]
That is, the municipal county dictionary is prepared for about 50 prefectures. For example, as shown in FIG. 5, the municipal county dictionary in which the prefecture is “Aichi prefecture” includes the cities belonging to the Aichi prefecture. The name of the ward and the recognition word are stored. Further, the town / village character dictionary is prepared for about 50 prefectures and for each municipal district of a specific prefecture. For example, as shown in FIG. 6, a name of a town and village character belonging to Kariya city in Aichi prefecture and a recognition word are stored. Note that the chome dictionary shown in FIG. 7 and the address dictionary shown in FIG. 8 do not belong to a specific place because they have names composed of numerals.
[0027]
Further, as shown in FIG. 9, the recognition dictionary storage unit 104 has a command code dictionary that stores a recognition word different from a name of a prefecture, a city, a county, or the like and a command code in association with each other. In this command code dictionary, for example, when the recognition word is “return”, the corresponding command code “C0001” is extracted. This command code is a code that can be recognized by the function execution unit 110 of the route guidance unit 11 described later.
[0028]
After extracting the name of the prefecture from the recognition dictionary storage unit 104, the recognition program processing unit 102 automatically switches the recognition dictionary to be collated / extracted next time to the city / ward / county dictionary corresponding to the name of the prefecture. . For example, if the prefecture name is "Aichi prefecture", the recognition dictionary to be collated / extracted next time is automatically switched to the city / ward / county dictionary corresponding to Aichi prefecture. This is the same even after matching / extracting the name of the city / ward / county. The recognition dictionary to be checked / extracted next time is automatically switched to the town / village dictionary corresponding to the name of the city / ward / county. However, the command code dictionary in FIG. 9 does not correspond to the above-described switching of the recognition dictionary, and is a recognition dictionary that is always a target of collation / extraction.
[0029]
Then, the recognition program processing unit 102 outputs to the route guidance unit 11 a signal corresponding to the name of the prefecture, city, county or the like obtained by the above processing. For example, when the voice of “Aichiken” of the prefecture is input, a signal corresponding to the name of the prefecture “Aichi” is transmitted.
[0030]
Next, the route guidance unit 11 of the car navigation device 1 will be described with reference to FIG. As shown in the figure, the route guidance unit 11 has a function execution unit 110. The function execution unit 110 executes a function of displaying a road map around the current location, a point search function by inputting an address, and the like. For example, in the function of displaying a road map around the current position, the vehicle position / vehicle direction calculation unit 12 receives a vehicle position / vehicle traveling direction signal, reads map data around the vehicle position from the map database 9, and outputs an image signal 15. And display it on the display device 4.
[0031]
On the other hand, in the point search function by inputting an address, the names of prefectures, municipalities, and the like transmitted from the voice recognition unit 10 are displayed on the display device 4, and the names of prefectures, municipalities, towns, villages, streets, and addresses are displayed. For example, a mark of a point corresponding to an address having a name such as a name and a road map around the mark are displayed on the display device 4.
[0032]
For example, when a signal corresponding to the name of a prefecture “Aichi” is received, the name of the prefecture is displayed on the display device 4. When a command code C0001 different from the name of the prefecture, city, or county is received, only the name located at the lowest level among addresses stored and displayed in a hierarchically connected manner is deleted. I do. Further, when the command code C0001 is received twice consecutively, all the names of the addresses stored and displayed are collectively erased. Note that, when the command code C0001 is received, a guidance sound, a message, or the like for executing display deletion may be notified via the speaker 5.
[0033]
Further, for example, when all the names of the addresses “1-1, Showa-cho, Kariya-shi, Aichi” have been received, the coordinates of the point corresponding to the received address are extracted from the map database 9 and further extracted. It reads out map information and road information around the coordinates. Then, the read information is converted into an image signal, and the display device 4 displays a mark of a point corresponding to the address and a road map around the mark.
[0034]
Next, in the above-described car navigation device 1, a process of a point search function for performing a point search based on an address by voice input will be described with reference to flowcharts of FIGS. 10 to 13 and a display image diagram of the display device 4 of FIG. explain. Note that, as a specific example, the description will be made on the assumption that the user inputs an address of “1-chome, Showa-cho, Kariya-shi, Aichi” and then re-enters the name from the prefecture name.
[0035]
First, in step S1 of FIG. 10, the standby state is continued until the talk switch 3 is pressed by the user, and when the talk switch 3 is pressed, the process proceeds to step S2. In step S2, the voice recognition unit 10 switches to the input mode, and enters a state of receiving a voice input. In addition, the voice recognition unit 10 switches the recognition dictionary so that the matching of the recognition word to be executed from now on and the extraction of the name of the prefecture, city, district, etc. corresponding to the recognition word are performed based on the prefecture dictionary. .
[0036]
Next, the speech recognition processing in step S3 will be described with reference to the flowcharts in FIGS. First, in step S30 shown in FIG. 11, the variable "Back" is initialized to zero. The variable “Back” is incremented by 1 each time the recognition word corresponding to the command code C0001 is collated / extracted. That is, by referring to the variable “Back”, the command code C0001 is output twice. It can be determined whether or not they have been continuously extracted.
[0037]
In a step S31, it is determined whether or not the user has spoken. If there is a speech, the process proceeds to the next step. If there is no speech, the process enters a standby state until there is a speech. In the present embodiment, it is assumed that there is an utterance “Aichiken”.
[0038]
In step S32, the utterance of the user is analyzed, and the analyzed recognized word reading is compared with the recognized words stored in the prefectural dictionary. Then, the most likely recognized word and the name of the prefecture corresponding to the recognized word are extracted. In the present embodiment, it is assumed that a prefecture name “Aichi Prefecture” corresponding to the recognition word “Aichiken” has been extracted.
[0039]
In step S33, it is determined in which hierarchy the recognition dictionary that has been collated / extracted in step S32 is one of the prefectural dictionary, the city / ward county dictionary, the town / village dictionary, the chome dictionary, and the address dictionary. In the present embodiment, it is determined that the recognition dictionary is the first hierarchy recognition dictionary.
[0040]
Further, in step S33, the recognition dictionary to be collated / extracted next time is switched to a recognition dictionary located at a lower hierarchy of the current recognition dictionary and corresponding to the currently extracted name. In the present embodiment, the recognition dictionary is switched to the recognition dictionary of the city, ward and county corresponding to the name of "Aichi prefecture" in the second hierarchy, which is lower than the first hierarchy. If the current recognition dictionary is at the fifth level, there is no recognition dictionary located at a lower level, and therefore, the process of switching the recognition dictionary is not executed.
[0041]
In step S34, the hierarchy of the recognition dictionary determined in step S33 is stored. In the present embodiment, the variable “kaisou” is set to 1 because it is the first-level recognition dictionary.
[0042]
In step S35, the name extracted in step S32 is displayed on the display device 4. In the present embodiment, as shown in FIG. 14, the display content 20 "Aichi Prefecture" is displayed.
[0043]
The step S36 decides whether or not the variable "kaisou" is 5, or whether a command indicating the end of the utterance such as "set" or "new set" is uttered, and the variable "kaisou" is 5. Alternatively, when a command indicating the end of the utterance is uttered, it is determined that the input of the names of the first to fifth layers has been completed, and the voice recognition processing ends. If this is not the case, the process proceeds to step S37. In the present embodiment, since the variable “kaisou” is 1, the process proceeds to step S37.
[0044]
The determination in step S37 is the same as that in step S31, and a description thereof will not be repeated. In step S38, the utterance of the user is analyzed, the analyzed recognized word reading is compared with the recognized word stored in the city / ward county dictionary, and a city / ward / county name corresponding to the recognized word is extracted. In addition, even if the analyzed recognized word reading is compared with the recognized word in the city / ward county dictionary, if the most reliable recognized word cannot be selected, the matching from the recognized word in the command code dictionary is attempted. Then, when the recognized word is collated from the command code dictionary, a command code corresponding to the recognized word is extracted. In the present embodiment, it is assumed that the city name “Kariya” corresponding to the recognition word “Kariyashi” is extracted.
[0045]
A step S39 decides whether or not the command code C0001 corresponding to the recognition word "return" has been extracted in the step S38. Here, if the command code C0001 is extracted, the process proceeds to step S50 in FIG. 13, and if not, the process proceeds to step S40 in FIG. In the present embodiment, since the city name “Kariya” has been extracted, the process proceeds to step S40.
[0046]
In step S40 shown in FIG. 12, the variable “Back” is set to zero. Here, the reason why the variable “Back” is set to zero is to switch the processing between the case where the user utters “return” twice consecutively and the case where the user utters “return” plural times although it is not twice consecutive. That's why. In step S40, since the command code C0001 corresponding to the recognized word "return" was not extracted in step S39, zero is set to the variable "Back".
[0047]
In step S41, it is determined in which hierarchy the recognition dictionary that has been collated / extracted in step S38 is a recognition dictionary of a prefectural dictionary, a municipal county dictionary, a town / village dictionary, a street dictionary, and an address dictionary. In the present embodiment, it is determined that the recognition dictionary is the second hierarchy recognition dictionary.
[0048]
Further, in step S41, the recognition dictionary to be collated / extracted next time is switched to a recognition dictionary located at a lower hierarchy of the current recognition dictionary and corresponding to the currently extracted name. In the present embodiment, the recognition dictionary is switched to the town / village character recognition dictionary corresponding to the name of “Kariya City, Aichi Prefecture” in the third hierarchy, which is lower than the second hierarchy. If the current recognition dictionary is at the fifth level, there is no recognition dictionary located at a lower level, and therefore, the process of switching the recognition dictionary is not executed.
[0049]
In step S42, the hierarchy of the recognition dictionary determined in step S41 is stored. In the present embodiment, since the recognition dictionary is a second-level recognition dictionary, the variable “kaisou” is set to 2 and stored.
[0050]
In step S43, the name extracted in step S38 is stored and displayed on the display device 4. In the present embodiment, as shown in FIG. 14, display content 21 "Kariya City, Aichi Prefecture" is displayed. Note that a name located at a lower hierarchy, such as the display content 21, is displayed on the display device 4 by being connected to an upper hierarchy. After that, the processing shifts to the step S36.
[0051]
In step S36, as described above, it is determined whether the variable "kaisou" is 5, or whether a command indicating the end of the utterance, such as "set" or "new set", has been uttered. In the present embodiment, since the variable “kaisou” is 2, the process proceeds to step S37.
[0052]
In step S37, in the present embodiment, the process proceeds assuming that there is an utterance of "showacho". In step S38, the analyzed recognized word reading is collated with the recognized word stored in the town and village character dictionary, and the recognized word is recognized. Extract the corresponding town character name. In the present embodiment, it is assumed that the town name “Showa-cho” corresponding to the recognition word “showacho” has been extracted. In step S39, the process proceeds to step S40 since the town name "Showa-cho" has been extracted in step S38.
[0053]
In step S40, the variable “Back” is set to zero again. In step S41, it is determined in step S38 that the recognition dictionary is the third-layer recognition dictionary, and the next recognition dictionary to be compared / extracted is set to the third-layer recognition dictionary. The mode is switched to the fourth level, that is, the lower level hierarchical dictionary. In step S42, in the present embodiment, since the recognition dictionary is the third hierarchy, the variable “kaisou” is set to 3 and stored. In step S43, the name extracted in step S38 is stored and stored in the display device 4. indicate. In the present embodiment, as shown in FIG. 14, display content 22 "Showa-cho, Kariya-shi, Aichi" is displayed. After that, the processing shifts to the step S36.
[0054]
In step S36, as described above, it is determined whether the variable "kaisou" is 5, or whether a command indicating the end of the utterance, such as "set" or "new set", has been uttered. In the present embodiment, since the variable “kaisou” is 3, the process proceeds to step S37.
[0055]
In step S37, in the present embodiment, the process proceeds assuming that there is an utterance of "Ichichome". In step S38, the analyzed recognized word reading is collated with the recognized word stored in the chome dictionary, and the recognized word is matched. Extract the name of the chome to perform. In the present embodiment, it is assumed that a chome name “1 chome” corresponding to the recognition word “ichichome” has been extracted. In step S39, the process proceeds to step S40 since the street name "1-chome" is extracted in step S38.
[0056]
In a step S40, the variable “Back” is set to zero, and in a step S41, it is determined in the step S38 that the recognition dictionary is a fourth-layer recognition dictionary, and the next recognition / extraction dictionary to be compared / extracted is set in a lower layer of the fourth layer. Is switched to the fifth level address dictionary. In step S42, since the recognition dictionary is the fourth hierarchy in the present embodiment, the variable “kaisou” is set to 4 and stored, and in step S43, the name extracted in step S38 is displayed on the display device 4. . In the present embodiment, as shown in FIG. 14, display content 23 "1-chome, Showa-cho, Kariya-shi, Aichi" is displayed. After that, the processing shifts to the step S36.
[0057]
In step S36, as described above, it is determined whether the variable "kaisou" is 5, or whether a command indicating the end of the utterance, such as "set" or "new set", has been uttered. In the present embodiment, since the variable “kaisou” is 4, the process proceeds to step S37.
[0058]
In step S37, in the present embodiment, the process proceeds assuming that there is an utterance of "return". In step S38, the analyzed recognized word reading is compared with the recognized word stored in the address dictionary, and the address corresponding to the recognized word is read. Extract the name. However, the recognition word reading “return” of the present embodiment does not select the most probable recognition word even when collating from the recognition word in the address dictionary. Therefore, matching is attempted from the recognized words in the command code dictionary. Then, in the present embodiment, the command code C0001 corresponding to the recognition word “return” is extracted.
[0059]
A step S39 decides whether or not the command code C0001 corresponding to the recognition word "return" has been extracted in the step S38. In the present embodiment, since the command code C0001 has been extracted, the process proceeds to step S50 in FIG.
[0060]
In step S50 of FIG. 13, 1 is added to the variable "Back". In this embodiment, since the variable “Back” is set to zero in the immediately preceding step S40, the variable “Back” is set to 1 here. That is, since the command code C0001 has been extracted in the immediately preceding step S38, 1 is added to mean that the command code has been extracted once.
[0061]
A step S51 decides whether or not the variable "Back" is 2; if the variable "Back" is 2, the process shifts to a step S54; otherwise, the process goes to a step S52. Proceed with the process. In the present embodiment, since the variable “Back” is 1, the process proceeds to step S52.
[0062]
In step S52, 1 is subtracted from the variable "kaisou", and the recognition dictionary to be collated / extracted next time is switched to the recognition dictionary which is one level higher than the current recognition dictionary. In the present embodiment, the dictionary is switched to the town / village dictionary corresponding to “Kariya City, Aichi Prefecture” which is the third hierarchy which is one hierarchy higher than the fourth hierarchy and which is the first and second hierarchy.
[0063]
In step S53, the storage and display of the name located at the lowest level among the addresses displayed on the display device 4 are deleted. In the present embodiment, as shown in the display content 24 of FIG. 14, the display of the fourth hierarchical level “1” is deleted. Thereafter, the process proceeds to step S37.
[0064]
In step S37, in the present embodiment, the process proceeds assuming that there is an utterance "return" again by the user. In step S38, the analyzed recognized word reading is collated with the recognized word stored in the town and village character dictionary, and the recognition is performed. Extract the town / village name corresponding to the word. However, the recognition word reading “return” of the present embodiment does not select the most probable recognition word even if it is collated from the recognition word of the Machimura dictionary. Therefore, matching is attempted from the recognized words in the command code dictionary. Then, in the present embodiment, the command code C0001 corresponding to the recognition word “return” is extracted.
[0065]
In the step S39, since the command code C0001 is extracted, the process shifts to a step S50. In the step S50, 1 is added to the variable "Back". In the present embodiment, since 1 is added to the variable “Back” in the immediately preceding step S50, the variable “Back” is 2 here. That is, it means that the command code C0001 is extracted twice consecutively. In step S51, since the variable “Back” is 2, the process proceeds to step S54.
[0066]
In step S54, of the addresses displayed on the display device 4, the storage and display of names located at all levels are deleted. In the present embodiment, as shown in the display content 25 of FIG. 14, all the names displayed so far are collectively deleted. Thereafter, the process proceeds to step S30, and the process is executed again from the input of the prefecture name.
[0067]
Note that when all the names from the first to fifth layers have been input, or even when no input has been made to the fifth layer, a command such as “set” or “new set” that means the end of the utterance is used. If uttered, the process proceeds from step S36 to step S4 in FIG. Then, in step S4, the coordinates of a point corresponding to the input address are extracted from the map database 9, and map information and road information around the extracted coordinates are read.
[0068]
Then, in step S5, the information read in step S4 is converted into an image signal, and the display device 4 displays a mark of a point corresponding to the address and a road map around the mark.
[0069]
As described above, when the speech recognition device of the present invention recognizes "return" twice in succession, it erases the display of the addresses that are connected hierarchically and displayed collectively. Thus, when inputting again from the prefecture during the input, the user can erase the display of the address in a lump by uttering “return” twice consecutively. As a result, the display can be easily erased without repeatedly saying “return”.
[0070]
In the present embodiment, “return” is set as a recognition word for the command code C0001, but the present invention is not limited to this. Furthermore, in the command code dictionary, for example, recognition words such as “erase” and “return to the beginning” and command codes C0002 for these recognition words are provided. When the command codes C0002 are extracted, The display of the addresses displayed jointly may be collectively deleted.
[0071]
Further, the scope of application of the present invention is not limited to the voice address input function of the car navigation device, and may be, for example, a function of voice inputting a telephone number or the like that is hierarchically indicated as in the case of an address. Applicable.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a schematic configuration of a car navigation device 1 according to the present embodiment.
FIG. 2 is a block diagram illustrating a configuration of a speech recognition unit 10 according to the embodiment.
FIG. 3 is a block diagram illustrating a configuration of a route guide unit 11 according to the present embodiment.
FIG. 4 is a diagram showing a prefectural dictionary according to the embodiment.
FIG. 5 is a diagram showing a dictionary of municipalities belonging to Aichi prefecture according to the present embodiment.
FIG. 6 is a diagram showing a dictionary of town and village characters belonging to Kariya city, Aichi prefecture, according to the present embodiment.
FIG. 7 is a diagram showing a chome dictionary according to the embodiment;
FIG. 8 is a diagram showing an address dictionary according to the embodiment.
FIG. 9 is a diagram showing a command code dictionary according to the embodiment.
FIG. 10 is a flowchart showing a flow of overall processing of the car navigation device 1 according to the embodiment.
FIG. 11 is a flowchart illustrating an overall flow of a voice recognition process according to the embodiment.
FIG. 12 is a flowchart illustrating a partial flow of a voice recognition process according to the embodiment;
FIG. 13 is a flowchart showing a flow of a display erasing process of a voice recognition process according to the embodiment.
FIG. 14 is a diagram showing a display image of the display device 4 according to the present embodiment.
FIG. 15 is a diagram showing a result of voice input according to the related art.
[Explanation of symbols]
1 Car navigation system
2 microphone
3 Talk switch
4 Display device
5 Speaker
6 GPS receiver
7 Vehicle speed sensor
8 Yaw rate sensor
9 map database
10 Voice Recognition Unit
11 route guidance
12 Vehicle position / vehicle direction calculation unit

Claims (3)

Voice input means for inputting voice;
Voice recognition means for recognizing a user's uttered voice input to the voice input means,
When a voice is input by the user, the recognized utterance contents are classified and stored in a plurality of layers, and when a correction command for correcting the utterance contents is input, the recognized utterance contents are classified and stored in the lowest layer. Control means for erasing a part of the utterance content,
The voice control device, wherein when the correction command is input twice consecutively, the control unit deletes all of the utterance contents stored in a plurality of layers. The voice control device according to claim 1, wherein the utterance content classified and stored in a plurality of layers is at least an address. 3. The voice control device according to claim 2, wherein the addresses are classified into at least five levels of prefectures, municipalities, towns and villages, streets and addresses.

Images