JP2009266129A - Data search device, navigation apparatus and data search method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data search device that can refine search results by a different search category in a navigation apparatus capable of search by search category. <P>SOLUTION: The data search device 50, which searches for data by search category on a search word detected by speech recognition, has a search category discrimination means 22 for discriminating the search category of the search word, a search word storage means 23 for storing each search word in association with the search category in time series of utterance, and a search means 22 for searching on the search word for data of the search category associated with the search word and then searching the search results on the next stored search word for data of the search category associated with the search word. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a data search apparatus that searches data by recognizing a voice uttered by a user, and more particularly to a data search apparatus, a navigation apparatus, and a data search method that search data for each category.

  2. Description of the Related Art Navigation devices that can perform more convenient voice operations during driving than manual operations are known. In voice operation, the destination is set in the navigation device by recognizing the voice, but according to the voice recognition feature that the recognition rate decreases as the number of recognition targets increases, a hierarchical dictionary is created and addresses are narrowed down by hierarchy. In many cases, a search method is used.

  FIG. 9 is a diagram for explaining a procedure of a narrowing search by conventional voice recognition. For example, when searching for an address “Minato-ku Daiba” in Tokyo, a dictionary of a hierarchy of addresses in prefectures is searched based on the voice “Tokyo” uttered by the user, and Tokyo hits. When Tokyo hits, the navigation device searches for a hierarchy whose address includes Tokyo.

  Next, when the user utters “Minato Ward”, since Minato Ward is hit from an address including Tokyo, the navigation apparatus searches for a hierarchy including “Tokyo Minato Ward” in the address. Similarly, when the user utters “Daiba”, “Daiba” is hit from the address including “Minato-ku, Tokyo”, so the navigation device searches for a hierarchy including “Daiba, Tokyo” in the address. .

In this way, since a desired position is narrowed down in a hierarchical manner, if voice operation is interrupted during voice input, the user may need to input an address again from the highest hierarchy, resulting in reduced operability. There was a case. In view of this, there has been proposed a navigation device that outputs a voice conversation history when voice operation is interrupted (see, for example, Patent Documents 1 and 2). In the search method described in Patent Document 1, the content of the speech input before the interruption is output when the voice input is resumed, and in the search method described in Patent Document 2, the input from the lower level than the already input hierarchy is performed To facilitate hierarchical searching.
Japanese Patent No. 3505982 JP 2004-240244 A

  However, the search method described in Patent Document 1 or 2 has a problem that a desired position cannot be searched unless the address is known. For example, if the user knows that a desired position exists in Shibuya Ward instead of Minato Ward, an operation of “return to the previous screen and narrow down with another address” can be performed. However, in such a method, if the search word (address in this case) of the destination is not known, the search is repeated many times, and there is a problem that the search target cannot be finally specified.

  In this regard, in navigation devices, it is often possible to search by name or telephone number in addition to an address when searching for a destination or the like. If the search by address is not possible, the user performs a search by “name” of the destination, for example, but the user returns to the selection screen of the search category (address search, name search, telephone number) and then searches by name. Execute. In FIG. 9, the search is performed by saying the name “Bia”. In other words, some inefficient operations have to be performed, such as a name search is not possible unless the address search is completed and the search category is selected.

  In this case, if the input of the address and the input of the name can be performed by one GUI (Graphical User Interface), it is possible to narrow down and search by both the address and the name. However, if the operation is not based on voice recognition, it is difficult to implement a GUI formed on a relatively small screen of the navigation device because the user's operation becomes complicated. Further, in the operation based on voice recognition, it has not been considered to narrow down search results across different search categories.

  In view of the above-described problems, the present invention provides a navigation device that searches for each search category, a data search device, a navigation device, and data that can be further refined by another search category using the search results so far The purpose is to provide a search method.

  In view of the above-described problems, the present invention provides a search category discrimination means for discriminating a search category of a search word in a data search apparatus that searches data for each search category using a search word detected by speech recognition. Search word storage means for associating and storing a search category for each search word in time series, and searching for search category data associated with the search word using the search word, and the search result And a search means for searching the search category data associated with the next stored search word using the search word.

  According to the present invention, since the search category can be made variable for each search word, a narrow search can be performed across a plurality of search categories. In addition, the search category is automatically identified from the voice, so the GUI is not complicated.

  Further, according to one aspect of the present invention, there is provided search category updating means for detecting a search category switching request by voice recognition and switching a search category associated with a search word.

  According to the present invention, since the search category can be switched by voice, for example, it is not necessary to return to the search category switching screen, and the search can be narrowed down using the search results so far.

  Further, in one embodiment of the present invention, the search category update means stores the search word after switching the search category in association with the search category before and after switching, and the search means stores data of the search category after switching. Is searched using a search word.

  According to the present invention, since the search is performed in the search category after switching, the search can be performed in the search category that can be expected to obtain the search result desired by the user with the smallest number of searches.

  In one embodiment of the present invention, the search category of the upper search word uttered earlier is an address, and the search category of the lower search word uttered after the search word is switched from an address to a name. In this case, the search means is characterized in that the search result by the upper search word is further narrowed down by the lower search word.

  According to the present invention, it is possible to narrow down and search a search result by name to some extent without returning to the search category search screen.

  It is a navigation device that searches for each search category, and a data search device, navigation device, and data search method capable of further narrowing down search by another search category using the search results so far can be provided.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

[Outline of Navigation Device 100]
FIG. 1 is a diagram for explaining an outline of a narrowing search by the navigation device 100 of the present embodiment. As shown in FIG. 9, conventionally, the search can be advanced only in the vertical direction on the left side of FIG. 1, but the navigation apparatus 100 of the present embodiment uses another result of narrowing down according to the search category so far. It is possible to further narrow down the search by the search category.

  The search category includes, for example, a search by address, a search by name, and a search by telephone number, but can be similarly applied even if other search categories such as a postal code are implemented. Note that in this embodiment, the user may be any person who speaks, but is, for example, a vehicle driver or a passenger.

In FIG. 1 (a), after inputting "Tokyo", the search result of address: "Tokyo" can be narrowed down by name: "Tesuni" by inputting "Tesuni". It is possible to search for “Tokyo Tesny Sea” and “Tokyo Tesny Land”.

In FIG. 1B, after inputting “Tokyo” and “Minato-ku”, the search result of address: “Tokyo” and “Minato-ku” is narrowed down by name: “Bia” by inputting “Bia”. You can search for “Ticket Beer Akasaka” and “Trade Beer Odaiba”.

  In FIG. 1C, after inputting “Tokyo”, “Minato-ku” and “Daiba”, the search category can be switched by inputting “Kirikae Meisho”. The user is notified that a plurality of search categories are identified for “Daiba”. As a result, the search category is switched from the address to “name”, and the search results for the address: “Tokyo” and “Minato-ku” can be narrowed down by name: “Daiba”, “Odaiba Beach” and “Trade Beer”. "Odaiba" can be searched. In this case, as will be described later, the final part of the search category ("address" of "Daiba" in FIG. 1C) is switched to the name.

  Thus, as shown in FIGS. 1A and 1B, the navigation device 100 of the present embodiment can perform a search by another search category while including the search results by the previous search category. Search results close to the intended destination can be obtained with a few steps. By using voice input for operation input, a user can be provided with many search categories without being restricted by a GUI (Graphical User Interface), and the GUI is not complicated, so that the operability is not lowered.

  Further, as shown in FIG. 1C, when a search by a plurality of search categories is possible (hereinafter referred to as a composite search), the user is notified, and the user can switch the search category. Therefore, a search by another category can be performed without returning to the search category selection screen, and a narrow search by another search category can be performed on the search results so far.

  FIG. 2 is a diagram illustrating a combined example of multiple searches. FIG. 2 shows a composite example of an address, a name, and a telephone number. When a place name representing an address is included in the name of the destination, a combined search of the address and name is possible. When the name includes a number that can be applied to a telephone number, a combined search of the name and the telephone number is possible. If the address contains a number that can be applied to a telephone number, a combined search of the address and telephone number is possible. When the address includes a number that can be applied to the name and telephone number of the destination, a combined search by the address, name and telephone number becomes possible.

[Functional Block Diagram of Navigation Device 100]
The navigation device 100 will be briefly described. FIG. 3 shows an example of a functional block diagram of the navigation device 100. The navigation device 100 is controlled by a navigation control unit 20, and the navigation control unit 20 includes a GPS receiver 11, a vehicle speed sensor 12, a gyro sensor 13, a database 14, via an in-vehicle LAN such as a CAN (Controller Area Network) or a dedicated line. The input unit 15 and the display 16 are connected.

  The GPS receiver 11 determines the position (latitude / longitude / altitude) of the vehicle based on the arrival times of radio waves from a plurality of GPS satellites. The vehicle speed sensor 12 measures, for example, the change in magnetic flux as pulses when the convex portions installed at regular intervals pass on the circumference of the rotor provided in each wheel of the vehicle, and the number of pulses per unit time Based on the above, the wheel speed is measured for each wheel. The vehicle speed can be obtained by multiplying the wheel speed by the outer diameter of the tire. The gyro sensor 13 is a vibration piece type gyro sensor formed by, for example, micromachining, and the vehicle rotates in the yawing direction or the pitching direction with respect to the road surface based on the distance of the vibration piece (between the electrodes) that changes due to the action of the Coriolis force. Detect the angular velocity of the hour. By integrating the detection result, the detection result is converted into an angle, that is, a traveling direction in the triaxial direction.

  The database 14 includes road map information 141 for forming a road network and various point information. The road map information 141 is a table-like database that associates nodes and links having position information corresponding to an actual road network. In the road map information 141, for example, the country is divided into meshes along latitude and longitude lines, and is handled by dividing them into primary meshes to tertiary meshes determined according to the degree of division. Since the position information of each section is determined for each mesh, a road network is formed by applying nodes and links whose positions are known to this section.

  The point information indicates information of a certain point where each of them is independent or related, such as an address point, a lot number point, a landmark, an office point, a telephone number point, and a zip code point. The point information is represented by points or icons (symbol marks) on the map, but each point is data having attribute information. For example, in the case of an office, various information such as the name, address, postal code, and telephone number of the office is registered.

  Therefore, it is possible to generate an address DB 142 in which only the address is extracted, a name DB 143 in which only the name is extracted, and a telephone number DB 144 in which only the telephone number is extracted. If the search hits in any one category, other information is also associated. Can be extracted.

  Since the point information is associated with the primary to tertiary meshes of the road map information 141 by geocoding, the point information is displayed on the road map as characters indicating icons and addresses. . The point information and the road map information 141 can be referred to each other. For example, when an address is input, the position on the map can be known, and the address can be displayed on the map.

  Each of the road map information 141 or the point information arranged on the map as an icon or character is handled as one layer, and the presence or absence of drawing can be controlled for each layer. As described above, the address DB 142 is hierarchized into prefectures, municipalities within each prefecture, place names within the municipalities, and addresses within the place names. The database 14 may be stored when the vehicle is shipped or the navigation device 100 is shipped, or after the navigation device 100 is mounted on the vehicle, a part or all of the database 14 is downloaded from a predetermined server and stored. May be.

  The input unit 15 serves as a user interface for inputting operation information for the user to operate the navigation device 100. For example, a push button type keyboard, a remote controller, a touch panel formed on the display 16, and a voice recognition unit 150 for inputting a voice uttered by a user. The navigation apparatus 100 according to the present embodiment will be described mainly with respect to an operation mode using voice recognized by the voice recognition unit 150.

  The display 16 is a liquid crystal display, a head-up display, or the like, and is used to switch and display the above-described road map information 141, point information, or a television image. For example, a speaker is provided that outputs a voice message instructing a right or left turn in synchronization with route guidance displayed on the display 16 before the intersection.

  The navigation control unit 20 is a computer having a CPU, a ROM, a RAM, a nonvolatile memory, a CAN communication unit, an input / output interface, and the like. The CPU executes a program or is realized by an arithmetic circuit such as an ASIC (Application Specific Integrated Circuit). A data search unit 50, a position detection unit 60, and a route search unit 70. The data search unit 50 will be described later.

  The position detection unit 60 accumulates the travel distance detected by the vehicle speed sensor 12 in the direction detected by the gyro sensor 13 with the vehicle position information detected by the GPS receiver 11 as a starting point, thereby accurately determining the position of the vehicle. presume. The route search unit 70 searches for a route from the current position to the destination acquired by the data search unit 50, for example. Extract links that can be routes to the destination, replace each link with a cost according to link length, left / right turn, width, traffic restrictions, etc., and route the route with the smallest integrated value of these costs to the destination And decide.

[Data Search Unit 50]
The data search unit 50 will be described with reference to the functional block diagram of FIG. Since a part of the database 14 is referred to, the database 14 is included in the data search unit 50 in FIG. First, the voice recognition unit 150 will be described. A microphone is installed at the center of the steering wheel, the instrument panel, the driver side A pillar, etc., for example, only when the driver presses the button, the sound is collected from the driver's mouth and converted into an electrical signal, Input to the voice recognition unit 150.

  The voice recognition unit 150 extracts a feature amount effective for voice recognition by, for example, Fourier transforming the electric signal. This feature value is obtained by extracting a sound effective for human hearing to understand voice information.

  The speech recognition unit 150 refers to the acoustic model DB that indicates the correspondence between the feature quantity and the phonetic symbol based on the feature quantity obtained by the acoustic analysis, and a series of candidates that may correspond to the feature quantity. Extract phonetic symbols. There are usually multiple phonetic symbols as candidates. Therefore, the speech recognition unit 150 refers to a word DB that stores words that can be recognized in association with phonetic symbols, and extracts words that are associated with a plurality of phonetic symbols. Therefore, words that are not registered (does not exist) are not extracted even in sound information with similar pronunciation.

  This word (all when there are a plurality of words) is input to the keyword discrimination unit 21. The keyword discriminating unit 21 extracts only words that can be used in the navigation device 100 and sends them to the information / search processing unit 22. That is, since the navigation device 100 excludes words that are used infrequently, the recognition rate can be improved.

  The information / search processing unit 22 determines whether the input word is a search word in the database 14 or a system operation word for operating the system. The information / search processing unit 22 stores a table in which system operation words are registered, and determines that the word is a system operation word if there is a word that matches or is similar to the input word. In the case of a system operation word, the information / search processing unit 22 sends the system operation word to the system operation control unit 24.

  The system operation control unit 24 reads the operation information stored in association with the system operation word, and requests a program or the like for processing corresponding to the operation information. As a result, an operation menu can be displayed on the display 16 or an operation such as re-searching the route to the destination can be requested to the navigation device 100.

  Similarly, the information / search processing unit 22 stores a table in which candidate words that may be candidates for a search word are registered, and determines that a search word is found if there is a word that matches or is similar to the input word. To do. Note that words other than system operation words may be determined as search words.

  Then, the information / search processing unit 22 determines the search category of the database 14 based on the search word. The search category determines which of the address DB 142, the name DB 143, or the telephone number DB 144 is to be searched. When searching the address DB 142, the search category can be determined up to the hierarchy. A search category is stored in association with a part or format of the search word, and if the search word is, for example, a numeric phone number expression, the phone number DB 144 is assigned to the search category, “XX prefecture”, “○ If there is an address expression such as “KU (ku)” or “XX (city)”, the address DB 142 is determined as a search category, and otherwise, the landmark 143 is determined as a search category.

  The information / search processing unit 22 updates / references the search word storage device 23. The search word storage device 23 stores a search history of search words. FIG. 5 shows an example of a search history of search words stored in the search word storage device 23. The search word storage device 23 stores search words in the order in which the user speaks. A search category is stored in association with each search word. For example, FIG. 5A is a search history showing a case where the address DB 142 is searched in hierarchical order, which is stored in order of “Tokyo”, “Minatoku”, “Daiba”, and each corresponds to a search category of “address”. It is attached.

  In the determined search category, the information / search processing unit 22 searches the database 14 using the search word determined to be valid from the current search word and the previous search word. Whether or not the search word is determined to be valid is determined to be valid unless, for example, the user inputs “cancel” or “redo” operation until then. Therefore, since the search can be performed using the series of search words up to the previous time and the current search word, the search results can be narrowed across the search categories even if the search categories are switched.

  The information / search processing unit 22 stores the current search word in the search word storage device 23 in association with the end of the previous search word. The search result searched by the information / search processing unit 22 is input to the drawing unit 25. When search by a plurality of search categories is possible, the information / search processing unit 22 inputs search category switchable information for notifying that there are a plurality of search categories to the drawing unit 25.

  FIG. 5B shows an example of a search history stored in the search word storage device 23 in which a plurality of search categories are associated with one search word. A search category of “address” and “name” is associated with the search word “Minato”.

[Operation Procedure of Data Search Unit 50]
FIG. 6 is a flowchart showing the operation procedure of the data search unit 50. The operation procedure of FIG. 6 starts when the user presses a button provided at the center of the handle, for example.
・ S10 (voice input)
First, when the user utters a voice, the voice recognition unit 150 recognizes the voice and inputs some words to the keyword discrimination unit 21.
・ S20 (keyword judgment)
The voice uttered by the user is a combination of words having different lengths such as a series of numbers such as a telephone number when several words are continuously input like an address. The keyword discriminating unit 21 cuts the plurality of words into words having an appropriate length to be recognized as a search word or a system operation word, and sends the words to the information / search processing unit 22 in order. For the segmentation, for example, using a dictionary for segmentation, matching words or expressions (numbers and other boundaries, “XX prefecture”, “XXX (city)”, “XXX (city)”, etc.) If there is, you can divide it.
S30 (determination of search word or system operation word)
The information / search processing unit 22 in which words are input in order determines whether each word is a system operation word or a search word. In the case of a system operation word (No in S <b> 30), the information / search processing unit 22 sends the system operation word to the system operation control unit 24.
・ S40 (Determination of search category)
If it is a search word (Yes in S30), the information / search processing unit 22 determines a search category (S40). As described above, if the search word includes a telephone number expression, the search category is determined to be a telephone number, and if the address expression is included, the search category is determined to be an address. For example, when the search word is “Minatoku”, the address expression “Minato Ward” is detected, and the search category is determined to be an address. The information / search processing unit 22 stores the search word and the search category in the search word storage device 23 in association with each other. As a result, as shown in FIG. 5A, the search word and the search category are associated with each other and stored in time series.

  Incidentally, it may be difficult to determine whether the search word is an address or a name. For example, if only “Minato” is recognized by voice in Minato Ward uttered by the user, there can be two search categories of “Name” “Minato Mirai” and “Address” Minato Ward. In this case, all of the information / search processing unit 22 determines the search category. In this case, search category switchable information is generated.

  As shown in FIG. 5B, the information / search processing unit 22 stores a plurality of search categories in association with one search word in the search word storage device 23. Even when a plurality of search categories are registered, it is possible to execute simultaneous parallel processing in the step of specifying an input history, which will be described later, by handling the same order.

S50 (determination of search word referring to search history)
If the search category is determined for the input search word, it can be searched from the database 14, but in this embodiment, the search word including the search history is determined by referring to the search history.

  First, the information / search processing unit 22 refers to the search history stored in the search word storage device 23 and reads a search word stored from the beginning of the search. Note that the search word of the search history to be read is the shortest state of the search word input by the user. As a result, the result of the narrowing search based on the search history is brought close to the result intended by the user.

・ S60 (update search category)
The shortest state will be described with reference to FIG. The shortest state is a combination of search categories that can be expected to be obtained with the smallest number of search results desired by the user. The shortest state is obtained from a combination of search categories last input by the user, or is obtained by switching the search categories by the user and updating the search categories in the search word storage device 23.

  In FIG. 7 (a), “TOKYO”, “Minatoku” and “Daiba” are stored as search histories as search words. In the state of FIG. 7 (a), when the user inputs “Kiri-kae” or “Meiseko”, the search category is input by operating to switch “Daiba” as the address to “Daiba”. That's right. In this case, the information / search processing unit 22 updates the search category of the search word “Daiba” stored in the search word storage device 23 to “name”. In FIG. 7B, the line immediately below the search word is the shortest state of the search history. Such an update in the shortest state is also reflected in subsequent searches.

  In addition, when the search category is switched for a search word associated with a plurality of search categories, the search category is stored without being deleted. In FIG. 7C, the search category “address” and “name” are associated with the search word “Tokyo”, but the search category of “Tokyo” is switched to “address” and “name”. The information / search processing unit 22 does not delete the “address” associated with “Tokyo”. Thereby, it becomes possible for the user to go back the search word and switch the search category. In FIG. 7C, “Tokyo” is the name, “Minatoku” is the “address”, and “Daiba” is the “name”.

  When “1” is input after “Daiba”, a search category is associated with “1” as shown in FIG. Since a number such as “1” may be an address number or a telephone number, “1” is associated with a plurality of search categories. In this case, which of the shortest information is determined may be the same as that in the previous search category, or may be determined based on past user switching operations.

S70 (search using search words)
The information / search processing unit 22 searches the database 14 using the search word in the shortest state. The search procedure will be described with reference to the flowchart of FIG. The information / search processing unit 22 initializes the search number counter N (S601). This initialization is to set “1” in the search number counter N. The search frequency counter N is a counter that determines that a fixed search has been performed using all search words in the minimum state.

  The information / search processing unit 22 reads the search word indicated by the search number counter N from the search word storage device 23 (S602). Similarly, the search category associated with the search word indicated by the search number counter N is read (S603). Then, according to the search category, the database 14 is searched using the search word from any one of the address DB 142, the name DB 143, and the telephone number DB 144 (S604).

  If the search is performed using one search word, the information / search processing unit 22 increments the search number counter N by 1 (S605). Then, it is determined whether or not the search number counter N exceeds the number of search words (S606). If the number of search counter N does not exceed the number of search words (No in S606), the search word remains, so the processes in steps S602 to S604 are repeated for the next search word. A search result corresponding to each search using one search word is obtained, and a search result narrowed down across search categories is obtained by repeating the search with respect to this search result.

  Note that the search results until finally presented to the user can reduce the processing load by using index information or the like used for identification in the database 14.

  If the search count counter N exceeds the number of search words (Yes in S606), the search is completed using all the search words, and the process is terminated.

S80 (drawing search results)
The drawing unit 25 acquires index information as a search result, and reads the search information to which the index number is assigned from the database 14. For example, the index numbers are displayed on the display 16 in ascending order.

  As shown in FIG. 5B, when a plurality of search categories are associated with one search word, the information / search processing unit 22 indicates that search categories can be switched. Possible information and search categories that can be switched are input to the drawing unit 25. Thereby, the drawing unit 25 can inform the user that the search can be performed in another search category. For example, in FIG. 1C, “The search category of Daiba can be switched to a name” is displayed.

・ S90 (system operation)
If it is determined in step S30 that a system operation word has been input, the system operation control unit 24 reads out the operation information in accordance with the system operation word and executes a process corresponding to the operation information. For example, a series of processes necessary for the system operation word can be determined by storing in advance the process corresponding to the system operation word. In other words, a series of processes necessary for a single system operation word is executed.

  For example, in the case of processing that requires redrawing, the system operation control unit 24 inputs display information to be redisplayed and information for requesting redrawing to the drawing unit 25.

  In addition, when a system operation word requesting a change in the search category (for example, “Kiri-kae” or “Meiseki”) is input, the search category of the search word stored in the search word storage device 23 is changed. Requests to the information / search processing unit 22. In response to this request, the information / search processing unit 22 executes the search category update processing in step S60 to update the search category of the search word at the end stored in the search word storage device 23, for example. As a result, the shortest information is updated.

  If the search is performed after the update of the shortest information, for example, a further refinement search using the name as a search category is executed for some refinement searches for addresses. Accordingly, the user can search for another search category without performing an operation of returning to the search category setting screen, so that the usability is improved.

・ S100 (drawing operation screen)
The drawing unit 25 redraws the display 16 in accordance with the input from the system operation control unit 24 separately from the drawing of the search result. For example, search result display scrolling or transition to a menu screen is executed.

  As described above, the navigation device 100 according to the present embodiment can perform a narrow search across a plurality of search categories without complicating the GUI by operating by voice recognition. When words that can be searched in a plurality of search categories are input, the user can be notified of this, and the search results so far can be searched and narrowed down in another search category. In this case, it is not necessary to return to the search category selection screen, and an efficient search is possible.

It is a figure explaining the outline | summary of the narrowing-down search by a navigation apparatus. It is a figure which shows the compound example of multiple searches. It is a figure which shows an example of the functional block diagram of a navigation apparatus. It is an example of the functional block diagram of a data search part. It is a figure which shows an example of the search history of the word for a search memorize | stored in a search word memory | storage device. It is a flowchart figure which shows the operation | movement procedure of a data search part. It is a figure explaining the shortest state. It is a flowchart figure which shows the procedure which searches a database. It is a figure explaining the outline | summary of the narrowing-down search by the conventional speech recognition.

Explanation of symbols

DESCRIPTION OF SYMBOLS 14 Database 15 Input part 16 Display 21 Keyword discrimination | determination part 22 Information and search process part 23 Search word memory | storage device 24 System operation control part 25 Drawing part 50 Data search part 100 Navigation apparatus

Claims (8)

In a data search device for searching data for each search category using a search word detected by speech recognition,
Search category discrimination means for discriminating a search category of the search word;
Search word storage means for storing the search category in association with each search word in time series spoken,
The data of the search category associated with the search word is searched using the search word, the search result is the search result, and the search word stored next is associated with the search word Search means for searching the data of the search category using the search word;
A data search apparatus characterized by comprising: Search category update means for detecting the search category switching request by voice recognition and switching the search category associated with the search word;
The data search apparatus according to claim 1, further comprising: The search category update means stores the search word after switching the search category in association with the search category before switching and after switching.
The search means searches the data in the search category after switching using the search word.
The data search apparatus according to claim 2, wherein: When the search category of the upper search word uttered earlier is an address, and the search category of the lower search word uttered after the search word is switched from the address to the name,
The search means performs a search by further narrowing down the search result by the upper search word with the lower search word.
The data search apparatus according to claim 2, wherein: Drawing means for displaying a search result searched by the search means on a display device;
5. The data search device according to claim 1, wherein the data search device includes: If the search category determining means determines that one search word has a plurality of search categories, the drawing means displays on the display device that there are a plurality of search categories;
6. A data retrieval apparatus according to claim 5, wherein: The data search device according to any one of claims 1 to 6,
Position detecting means for detecting the position;
Search means for searching for a route from the position to the destination input by the data search device;
A navigation device comprising: In a data search method for searching data for each search category using a search word detected by speech recognition,
A search category determining unit determining a search category of the search word, and storing the search category in the search word storage unit in association with the search category for each of the search words in the time series spoken;
A search means searches the data of the search category associated with the search word using the search word, and corresponds to the search word that is the search result and is stored next. Searching for the data of the search category attached using the search word;
A data search method characterized by comprising:

Images