JP2007271901A - Navigation device and its address input method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a navigation device which is easy for a user to perform voice input of an address, and improves precision of a recognition result. <P>SOLUTION: The navigation device includes: a means for receiving voice input of the starting one character of a street name, a means of recognizing the received speech; a means for discriminating street names starting with the recognized character as candidates; a means for receiving speech input of the whole street name; and a means for recognizing the received speech by using the candidates. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a navigation device and an address input method thereof.

  Since an address is composed of a state name, a city name, and a street name, there is a navigation device that utters in this order, recognizes a voice, and identifies a point (see Patent Document 1). On the other hand, in consideration of the case where there are a large number of streets, up to the second character is first spoken regarding the street name. Specifically, as shown in FIG. 8, up to two characters of the street name are uttered and recognized. Then, the recognized two-character candidates are displayed. The user is made to select by voice, and a street starting with the selected two characters is set as a candidate. After that, all the street names are uttered, and the best matching street name is extracted from the street name candidates and used as a recognition result.

JP 2003-140682 A

  However, depending on the street name, it may be difficult to determine which character is the first character. For example, when “SOUTH ABC STREET” is uttered up to the second character, “S, O” is correct. However, since “S ABC ST” is often described, some users utter “S, A”. In this case, an appropriate candidate cannot be determined, and speech recognition is not accurately performed.

  SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and an object thereof is to improve the accuracy of a recognition result that is convenient for a user in voice input of an address.

  In order to solve the above problem, the present invention first accepts the first character of the street name.

  For example, the first aspect of the present invention is a navigation device that accepts speech input of the first character of a street name, means for recognizing the received speech, and candidates for street names starting with the recognized character. Means for accepting all speech inputs of street names, and means for recognizing the accepted speech using the candidates.

  The navigation device includes a storage unit that stores a street name, a unit that receives voice input of the first character of the street name, and a first character candidate calculation unit that recognizes the received voice and obtains a candidate for the first character. Street name extraction means for extracting a street name starting with the first character candidate obtained by the first character candidate calculation means from among the street names stored in the storage means, and a street name (full name) Street name accepting means for accepting voice input and means for extracting a street name that most closely matches the voice accepted by the street name accepting means from the street names extracted by the street name extracting means .

Further, the second of the present invention is a navigation device, a storage means for storing a street name for each state name and city name, a means for receiving voice input of the state name,
A state name identifying means for recognizing received speech to identify a state name, a means for receiving speech input of a city name, a city name identifying means for recognizing received speech and identifying a city name, and a street name first A means for accepting a voice input of one character, a character reliability calculating means for recognizing the accepted voice to obtain a reliability for each letter of the alphabet, and a character extraction for extracting a character having a reliability equal to or higher than a predetermined value Means, street name extracting means for extracting a street name beginning with the character extracted by the character extracting means from street names belonging to the city name specified by the city name specifying means, and a street that accepts speech input of the street name Of the street names extracted by the name accepting means and the street name extracting means, the voice is the best match with the speech accepted by the street name accepting means. And means for extracting and outputting treat name.

  An embodiment of the present invention will be described below with reference to the drawings.

  FIG. 1 is a schematic configuration diagram of an in-vehicle navigation device 100 to which an embodiment of the present invention is applied. As shown in the figure, the in-vehicle navigation device 100 includes an arithmetic processing unit 1, a display 2, a storage device 3, a voice input / output device 4, an input device 5, a wheel speed sensor 6, a gyro sensor 8, And a GPS (Global Positioning System) receiver 9.

  The arithmetic processing unit 1 is a central unit that performs various processes. For example, the current position is detected based on information output from the various sensors 6 to 8 and the GPS receiver 9. In addition, the voice input from the voice input / output device 4 is recognized, and the input content of the user is specified from the recognized word / phrase.

  The display 2 is a unit that displays graphics information generated by the arithmetic processing unit 1.

  The storage device 3 includes a storage medium such as a CD-ROM, DVD-ROM, HDD, or IC card. This storage medium stores map data. The storage device 3 stores an address database. FIG. 2 shows a configuration example of the address database 300. As shown in the figure, the address database 300 stores a state name 301, a city name 302, and a street name 303 in a hierarchical structure.

  Returning to FIG.

  The voice input / output device 4 includes a microphone, acquires voice spoken by the user, and transmits the voice to the arithmetic processing unit 1. Further, the message to the user generated by the arithmetic processing unit 1 is converted into an audio signal and output.

  The input device 5 is a unit that receives instructions from the user. The input device 5 includes a hard switch such as a scroll key and a scale change key, a joystick, a touch panel pasted on a display, and the like.

  The sensors 6 to 8 and the GPS receiver 9 are used for detecting the current location (vehicle position) by the vehicle-mounted navigation device 100.

  FIG. 3 is a functional block diagram of the arithmetic processing unit 1.

  As illustrated, the arithmetic processing unit 1 includes a user operation analysis unit 41, a voice recognition unit 42, a navigation processing unit 43, an information storage unit 44, a display processing unit 45, and a voice dictionary database 46. Yes.

  The user operation analysis unit 41 receives a request from the user input to the input device 5, analyzes the request content, and controls each unit of the arithmetic processing unit 1 so that processing corresponding to the request content is executed. To do. Further, each part of the arithmetic processing unit 1 is analyzed such that a user request (command) is analyzed from the words input to the voice input / output device 4 and recognized by the voice recognition unit 42, and processing corresponding to the request content is executed. To control.

  The voice recognition unit 42 acquires voice spoken by the user via the voice input / output device 4. Then, the acquired voice is recognized using the voice dictionary database 46, and a phrase (word) uttered by the user is specified. For example, when a user speaks an address (State, City, Street) in order to specify a point, the state name, city name, and street name are recognized from these voices. The existing technology can be applied to the speech recognition method for recognizing words from speech. For example, a method using DP (dynamic programming) matching or a method using HMM (Hidden Markov Model) can be applied. Note that the speech dictionary database 46 stores speech models necessary for speech recognition in association with words.

  The navigation processing unit 43 obtains the current position from the outputs of the sensors 6 to 8 and the GPS receiver 9, and searches for a recommended route connecting between two designated points (current location, destination). In addition, the recommended route is displayed on the display 2 to guide the route.

  The information storage unit 44 stores words and phrases recognized by the voice recognition unit 42 and information on points searched by the navigation processing unit 43.

  The display processing unit 45 generates a drawing command for the display 2. For example, a map drawing command is generated so as to draw marks such as roads, other map components, current location, destination, and arrows for recommended routes with a specified scale and drawing method.

  FIG. 4 is a diagram illustrating a hardware configuration example of the arithmetic processing unit 1.

  As illustrated, the arithmetic processing unit 1 has a configuration in which devices are connected by a bus 32. The arithmetic processing unit 1 includes a CPU (Central Processing Unit) 21 that executes various processes such as numerical calculation and control of each device, and a RAM (Random Access Memory) that stores map data, arithmetic data, and the like read from the storage device 3. ) 22, a ROM (Read Only Memory) 23 for storing programs and data, a DMA (Direct Memory Access) 24 for transferring data between the memories and between the memory and each device, and graphics drawing. In addition, a drawing controller 25 that performs display control, a video random access memory (VRAM) 26 that stores graphics image data, a color palette 27 that converts image data into RGB signals, and an A / D that converts analog signals into digital signals. Converter 28 and an SCI (Serial Communication Interface) that converts the serial signal into a parallel signal synchronized with the bus. ace) 29, a PIO (Parallel Input / Output) 30 that puts a parallel signal on the bus in synchronization with the bus, and a counter 31 that integrates the pulse signal.

  [Description of Operation] Next, the operation related to the speech recognition of the address of the vehicle-mounted navigation device 100 having the above-described configuration will be described with reference to FIGS.

  FIG. 5 is a flowchart showing the flow of address speech recognition.

  The voice recognition unit 42 starts this flow when receiving a voice input request for an address from the user via the input device 5.

  First, the voice recognition unit 42 outputs a message prompting the utterance of the state name such as “Please utter a state” through the voice input / output device 4. At this time, the voice recognition unit 42 may display this message on the display 2 as shown in the display screen (a) of FIG. 7 (S11).

  The voice recognition unit 42 acquires the voice uttered by the user via the voice input / output device 4. Then, it recognizes the voice spoken by the user and identifies the state name. Specifically, among the state names included in the state name 301 of the address database 300, the state name that most closely matches the input voice is specified as the recognition result (S12).

  Next, the voice recognition unit 42 outputs a message prompting the utterance of the city name, such as “Please utter a city” via the voice input / output device 4. At this time, the voice recognition unit 42 may display this message on the display 2 as shown in the display screen (b) of FIG. 7 (S13).

  The voice recognition unit 42 acquires the voice spoken by the user via the voice input / output device 4. And the voice which the user uttered is recognized and a city name is specified. Specifically, the voice recognition unit 42 refers to the address database 300 and extracts a city name 302 belonging to the state name 301 specified in S12. Then, a city name that most closely matches the input voice is identified as a recognition result from the extracted city names 302 (S14).

  Next, the voice recognition unit 42 outputs a message prompting the first character of the street name, such as “Please utter the first character of the street name” via the voice input / output device 4. . At this time, the voice recognition unit 42 may display this message on the display 2 as shown in the display screen (c) of FIG. 7 (S15).

  The voice recognition unit 42 acquires the voice spoken by the user via the voice input / output device 4. Then, the voice spoken by the user is recognized. Specifically, as shown in FIG. 6, the speech recognition unit 42 obtains the reliability 52 for each character 51 from the correlation between the speech uttered by the user and the speech model of the alphabetic character (S16).

  Next, the voice recognition unit 42 extracts characters whose reliability 52 is greater than or equal to a predetermined value. If there are a plurality of characters, a plurality of characters are extracted (S17).

  Next, the speech recognition unit 42 refers to the address database 300, extracts street names belonging to the city name 302 specified in S14, and begins with the characters extracted in S17 from the extracted street names. Is extracted (see FIG. 6). The extracted street name is set as a street name candidate (standby) (S18).

  Next, the voice recognition unit 42 outputs a message prompting all utterances of the street name, such as “Please utter the street name”, via the voice input / output device 4 by voice. At this time, the voice recognition unit 42 may display this message on the display 2 as shown in the display screen (d) of FIG. 7 (S19).

  The voice recognition unit 42 acquires the voice spoken by the user via the voice input / output device 4. Then, the voice spoken by the user is recognized, and the reliability is obtained for each candidate of the street name obtained in S18 from the correlation of the voice (S20).

  Finally, the speech recognition unit 42 extracts the street name with the highest reliability from the candidates. Then, as shown in the display screen (e) of FIG. 6, the result is displayed on the display 2 as a street name speech recognition result (S21).

  When receiving a confirmation request from the user, the voice recognition unit 42 confirms the address. Thereafter, the voice recognition unit 42 may accept an address. Then, a map around the specified address may be displayed on the display processing unit 45.

  The embodiment of the present invention has been described above.

  According to the above embodiment, the user is not required to speak only the first character before inputting all the street names by voice. Therefore, the user does not have to worry about the second character. For example, since “S ABC ST” is often described as “SOUTH ABC STREET”, it is difficult to determine whether the second character is “O” or “A”. Even in this case, according to the present embodiment, since only the first character needs to be uttered, the user is not lost.

  A further advantage of the present invention is that the user is not confused even if a second character of the street name contains a character other than the alphabet. For example, in the case of street names such as “O′CONNELL ST”, “L & M RANCH RD”, it may not be understood how to read the second character. If it is sufficient to input only the first character, the user can input utterances “O” and “L” without confusion.

  When all street names are uttered and voice recognition is performed (S20 in FIG. 5), the voice most closely matches the voice input from the candidates regardless of the reliability of the first character voice recognition performed previously. Street names are extracted and used as recognition results. For example, as a result of recognizing the first character, even if “N” is more reliable than “M”, the street name starting with “M” is more reliable when speaking all of the street names. If it is larger, the street name starting with “M” is the result of speech recognition. Therefore, the street name can be accurately identified without being influenced too much by the result of recognition of the first character while narrowing down candidates by voice recognition of the first character.

  The present invention is not limited to the above embodiment. The above embodiment can be variously modified within the scope of the technical idea of the present invention.

  For example, in the above-described embodiment, an example in which the present invention is applied to an in-vehicle navigation device has been described. However, the present invention may be applied to other devices that recognize voice of an address.

FIG. 1 is a schematic configuration diagram of an in-vehicle navigation device to which an embodiment of the present invention is applied. FIG. 2 is a configuration example of the address database. FIG. 3 is a diagram illustrating a functional configuration of the arithmetic processing unit 1. FIG. 4 is a diagram illustrating a hardware configuration of the arithmetic processing unit 1. FIG. 5 is a flowchart of address speech recognition. FIG. 6 is a diagram for explaining how street name candidates are determined. FIG. 7 is a transition example of the display screen. FIG. 8 is a diagram for explaining how to determine candidates for street names in the conventional example.

Explanation of symbols

100: In-vehicle navigation device,
DESCRIPTION OF SYMBOLS 1 ... Arithmetic processing part, 2 ... Display, 3 ... Memory | storage device, 4 ... Audio | voice output / input device, 5 ... Input device, 6 ... Wheel speed sensor, 8 ... Gyro, 9 ... GPS receiver, 21 ... CPU, 22 ... RAM 23 ... ROM, 24 ... DMA, 25 ... drawing controller, 26 ... VRAM, 27 ... color palette, 28 ... A / D converter, 29 ... SCI, 30 ... PIO, 31 ... counter,
DESCRIPTION OF SYMBOLS 41 ... User operation analysis part, 42 ... Voice recognition part, 43 ... Navigation processing part, 44 ... Information storage part, 45 ... Display processing part, 46 ... Voice dictionary database

Claims (4)

A navigation device,
Means to accept voice input of the first letter of the street name;
Means for recognizing the received voice,
A means to suggest street names that start with recognized characters,
A means of accepting all voice input of street names;
A navigation device comprising means for recognizing received speech using the candidate. The navigation device according to claim 1,
Storage means for storing street names;
Means to accept voice input of the first letter of the street name;
First character candidate calculating means for recognizing the received speech and obtaining a candidate for the first character;
Street name extraction means for extracting a street name starting from the first character candidate obtained by the first character candidate calculation means from the street names stored in the storage means;
Street name accepting means for accepting voice input of street names,
A navigation apparatus comprising: means for extracting a street name that most closely matches the voice received by the street name receiving means from the street names extracted by the street name extracting means. A navigation device,
Storage means for storing the street name for each state name and city name, means for receiving voice input of the state name,
A state name identifying means for recognizing the received voice and identifying the state name,
Means for receiving voice input of city names,
A city name identifying means for recognizing the received voice and identifying the city name,
Means to accept voice input of the first letter of the street name;
A character reliability calculation means for recognizing the received speech and calculating the reliability of each letter of the alphabet;
A character extracting means for extracting characters having a reliability equal to or higher than a predetermined value;
Street name extraction means for extracting a street name starting with the character extracted by the character extraction means from street names belonging to the city name specified by the city name specifying means;
Street name accepting means for accepting voice input of street names,
A navigation apparatus comprising: means for extracting and outputting a street name that most closely matches the voice received by the street name receiving means from the street names extracted by the street name extracting means. An address input method for a navigation device,
Accepting voice input for the first letter of the street name;
Recognizing the received speech,
Candidate street names starting with recognized characters,
Accepting all voice input of street names;
Recognizing received speech using the candidates, and an address input method for a navigation device.

Images