JP3771437B2 - Information exchange method and information exchange system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a technique for transmitting and collecting information to the Internet or the like with an in-vehicle terminal, and more specifically, an information exchange method for transmitting, receiving, and presenting information while ensuring the safety of driving of a driver, and Information exchange system.
[0002]
[Prior art]
Due to the development of the information distribution industry and mobile communication technology, in the social situation where a vast amount of information sources can be accessed anytime and anywhere in a car such as an automobile, not only passive information access such as radio and car navigation information, but also mobile phones, for example Media that allows two-way access, such as telephone and e-mail, has been introduced. And safety has become a problem there, and if it is a mobile phone, it is obliged to use a hands-free device together. In addition, text input means such as e-mails are transmitted by a text selection method, and can be used only when the vehicle is stopped in principle.
[0003]
On the other hand, in an i-mode compatible car navigation system, it is now possible to perform web browsing and mail transmission / reception via the car navigation interface. Speech recognition is mainly employed as an interface for a car navigation system, and a technique such as dynamically configuring a speech environment file to improve the recognition rate is used.
[0004]
[Problems to be solved by the invention]
In the prior art, there is a great limitation when transmitting information from within the vehicle. For example, there is a limited number of selected sentences when creating a mail transmission sentence, and there is a problem that it cannot be configured from a selected sentence depending on the contents of transmission. Further, in order to expand the range of speech recognition, it is necessary to increase the number of candidate words for speech recognition, which has a problem of a trade-off relationship with deterioration of the recognition rate. In addition, it is necessary to manually input important information such as position, speed, and time in information exchange in mobile communication, and there is a problem that the user wastes time.
[0005]
The present invention solves the above-mentioned problems in information transmission from the inside of a vehicle, and provides a voice recognition method, selected sentence structure, and automatic input support according to the content, thereby ensuring comfort while ensuring safety. The purpose is to improve accessibility to information.
[0006]
[Means for Solving the Problems]
In a system in which an in-vehicle terminal arbitrarily communicates with an information storage / provision center to exchange information, a content description language that supports a user in a vehicle to transmit information and a browser that analyzes and executes the content description language It has the main feature of having.
[0007]
In the present invention, the center side (server side) analyzes the information transmitted from the office / home terminal, the extraction means for extracting the key word from the analysis result, The main feature is that it has a registration means for registering in the content as a selection word.
[0008]
In addition, the present invention provides a candidate word creation in which the in-vehicle terminal side dynamically creates a speech recognition candidate word by using the speech recognition word and category described in the question sentence content transmitted from the center side. The main feature is having means.
[0009]
Further, according to the present invention, the in-vehicle terminal side transmits information extracted from various sensors without a user's hand in accordance with the automatic input format described in the contents of the question sentence transmitted from the center side. The main feature is having an automatic input processing means.
[0010]
Further, in the present invention, the in-vehicle terminal side has a microphone mounted on the vehicle, and communication means for transmitting voice or sound acquired therefrom to the center, and the center side stores storage means for storing the voice and sound. The main feature is that it has a recognition means for performing speech recognition and a distribution means for distributing the accumulated or recognized result. In addition, the main feature of the center is that if it is sound, it is stored and distributed with a link to the content.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a block diagram showing a configuration example of an information exchange system according to the present invention. The system is roughly divided into a center facility 100 as an information storage / providing center and a vehicle terminal facility 200 as an in-vehicle terminal. When the center facility 100 processes / stores information transmitted from the office / home or the vehicle terminal facility 200, which is omitted in FIG. 1, or transmits content to the vehicle terminal facility 200, the center facility 100 determines the type of the in-vehicle terminal. It has a function of transmitting contents in a corresponding description format. The vehicle terminal facility 200 analyzes the content sent from the center facility 100 and automatically transmits information extracted from various sensors or the like according to the content or transmits information input from the user. Have.
[0012]
An example of the opportunity for the center facility 100 to transmit content to the vehicle terminal facility 200 side is as follows.
1. Send in response to a request from the user.
2. Frequently asked questions are sent as a fixed sentence to the first when the user activates the in-vehicle terminal.
3. It detects the position information on the vehicle side and transmits it when it reaches a predetermined position.
4). An expiration date is set for content (such as a bulletin board) and sent to all users who have accessed the content so far.
[0013]
First, the configuration and operation of the center facility 100 will be described in detail. The center facility 100 is roughly divided into a transmission processing unit 110, an item extraction unit 120, a content database 130, an acoustic processing unit 140, and a communication unit 150. The center facility 100 can be connected to each terminal via the Internet 300 or a communication line, and can be accessed from an office / home or a terminal in the car. A user accessing from an office or home accesses content and exchanges information in the same manner as in a conventional Internet environment.
[0014]
The item extraction unit 120 receives data from a communication unit 150 and analyzes data input by a user who accesses from an office or home from the communication unit 150, extracts items and contents, and uses them as selection words when making a call from the vehicle. Register in the database 130 . The item extraction unit 120 includes a morphological analysis unit 121, an item related dictionary 122, a content related dictionary 123, an expiration date setting unit 124, a key extraction unit 125, a data creation unit 126, and a database registration unit 127.
[0015]
The morphological analysis unit 121 analyzes data input by a user (particularly a user accessing from an office or home). The key extraction unit 125 extracts words / phrases that can be items from the data analyzed by the morpheme analysis unit 121 with reference to the item relation dictionary 122. Further, the key extraction unit 125 refers to the content-related dictionary 123 from the words and phrases extracted as items, and lists and extracts words and phrases that are the contents of the items. The expiration date setting unit 124 extracts words / phrases that can become expiration dates from the data. The data creation unit 126 creates the extracted data including items, contents, and expiration dates, and the database registration unit 127 registers this in the content database 130.
[0016]
FIG. 2 shows a processing example in the item extraction unit 120. For example, in the case of a question sentence “How is the weather?”, “Weather” is extracted as an item, and {sunny, cloudy, rain, thunder, snow, etc.} is set as the value. When the contents of the question sentence are transmitted to the terminal, these values (return values) can be expected as answers to the question in the transmission data (response) from the in-vehicle terminal.
[0017]
The transmission processing unit 110 converts the description format of the content corresponding to the terminal type, and performs transmission processing for the corresponding terminal. The transmission processing unit 110 includes a request analysis unit 111 and a conversion unit 112.
[0018]
The request analysis unit 111 extracts the content transmission destination terminal type (home / office / in-vehicle) and correspondence status (complete correspondence / incomplete correspondence / non-correspondence) by User Agent, and notifies the conversion unit 112 of the extracted information. The conversion unit 112 acquires the content to be transmitted from the content database 130, converts the content into a format according to the type / support status of the terminal, and transmits it to the corresponding terminal from the communication unit 150.
[0019]
Here, the format of the content transmitted to the in-vehicle terminal needs to be transmitted in a format that can withstand use in a special environment such as an in-vehicle environment. FIG. 3 lists examples of features of the in-vehicle environment. In the present invention, the description language of the content to be transmitted to the in-vehicle terminal is upwardly compatible with HTML or XML, or a subset of HTML or XML, and the in-vehicle device such as automatic input of sensor information, speech synthesis and recognition on the in-vehicle terminal side. This is a description format that can be controlled.
[0020]
FIG. 4 shows a specific example of content transmitted to the in-vehicle terminal. Here, an HTML-based header description example is shown. Since this example is a question about the weather, certain information can be automatically obtained from sensor information such as wipers and lights. This request for automatic acquisition is also described in the header (for example, <INPUT Type = auto ...>). If the type of terminal is home / office, the header as shown in FIG. 4 is unnecessary and is omitted. In addition, even if it is an in-vehicle terminal, in the case of a terminal that does not support automatic input, the “<INPUT Type = auto ...>” portion is not necessary.
[0021]
As will be described later, the in-vehicle terminal (vehicle terminal equipment 200) that has received the data described in this way analyzes the data and creates an interface that needs it (or is necessary to satisfy the data request). Send data and get the return value. In the example of FIG. 4, the voice input interface and the automatic input interface correspond to this. The in-vehicle terminal aggregates the return values and transmits (transmits) them to the center facility side.
[0022]
The acoustic processing unit 140 receives voice / acoustic data transmitted from the in-vehicle terminal or the like from the communication unit 150, analyzes and accumulates the acoustic / sound acquired by the in-vehicle microphone or the like, and links with the content in the content database 130, When a content distribution request is received from the terminal, acoustic data related to the content is transmitted via the transmission processing unit 110 and the communication unit 150. The acoustic processing unit 140 includes a voice recognition unit 141 and a storage unit 142.
[0023]
The voice recognition unit 141 performs voice recognition processing on particularly voice data sent from the terminal, and registers the recognition result (text data) in the content database 130. The accumulating unit 142 accumulates audio data or acoustic data sent from the terminal and links to the content in the content database 130. When there is a content request from the terminal, the requested content is searched from the content database 130, and when there is audio / sound data linked to the content, the corresponding audio / sound data is found from the storage unit 142, and the content Provide with.
[0024]
Next, the configuration and operation of the vehicle terminal facility 200 will be described. The vehicle terminal equipment 200 is roughly divided into a communication unit 210, a browser device 220, various sensors 240, a voice recognition unit 250, a voice synthesis unit 260, and various user interfaces 270. The vehicle terminal equipment 200 can be connected to the Internet 300 or the private net including the center equipment 100 via the communication unit 210 to exchange various information.
[0025]
The browser device 220 analyzes the content sent from the center facility 100, notifies the user by an optimal notification method according to the environment, and sends information obtained from the various user interfaces 260 and the various sensors 240 to the center facility 100. Play a role to send to. The browser device 220 includes a decoder 221, a user input interface (I / F) 222, a speaker I / F 223, a display I / F 224, a voice recognition I / F 225, a voice synthesis I / F 226, a sensor I / F 227, a microphone I / F 228, An automatic input processing unit 229 and a calculation unit 230 are included.
[0026]
User input I / F 222, speaker I / F 223, display I / F 224, voice recognition I / F 225, voice synthesis I / F 226, sensor I / F 227, microphone I / F 228, various sensors 240, voice recognition unit 250, voice synthesis It controls the interface with the unit 260 and various user interfaces 270.
[0027]
The decoder 221 analyzes the content sent from the center facility 100, classifies each I / F so as to be easily processed by the calculation unit 230, and transmits it to the calculation unit. FIG. 5 shows a classification example of the HTML-based header description example shown in FIG. 4 into sensor I / F automatic input, speech recognition I / F, and speech synthesis I / F. The underlined portion is classified so as to notify each I / F.
[0028]
The calculation unit 230 assigns the classified content sent from the decoder 221 to each I / F and performs scheduling thereof. For example, in the case of the description example of FIG. 5, scheduling is performed by transmitting speech recognition data to the speech recognition unit 250 via the speech recognition I / F 225 to perform preprocessing, and then performing speech synthesis I / F 226. The voice synthesis unit 260 notifies the user of the synthesized voice “What is the weather?” And displays the content corresponding to the screen, prompts the user interface 270 to input the voice via the microphone I / F 228, and then the voice. This is to control the flow of notifying the recognition unit 250 to start speech recognition. Further, the automatic input processing unit 229 is notified of items for which automatic input is designated for the content. FIG. 6 shows a specific example of processing.
[0029]
The automatic input processing unit 229 of the browser device 220 inputs values acquired from the various sensors 240 for the automatic input processing items notified from the calculation unit 230. Here, the various sensors 240 sense the driving situation, the situation of the driver, and the like, and notify the value in response to a request from the sensor I / F 227 in the browser device 220. The sensing value includes not only a real-time value but also a history such as “average speed over the past hour”.
[0030]
The voice recognition unit 250 first extracts a candidate word for voice sound recognition for the user input request processing item notified from the calculation unit 230, then performs voice recognition of the input voice, and inputs the result to the user input request item. . The voice recognition unit 250 includes a recognition processing unit 251, a synonym dictionary 252, and a candidate word extraction / storage unit 253.
[0031]
Here, since the data input in the in-vehicle environment assumes voice recognition in consideration of safety, it is necessary to narrow down recognition candidate words in order to improve the recognition rate. For this purpose, it is necessary to select input instead of free text input. However, since the degree of freedom is greatly reduced when the selected sentence is input, the synonym dictionary 252 is used to increase the tolerance of the spoken word in order to compensate for it. Specifically, the number of candidate words is increased.
[0032]
The candidate word extraction / storage unit 253 performs preprocessing of candidate word extraction from the speech recognition data sent from the calculation unit 230 of the browser device 220 using the synonym dictionary 252. For example, the synonym of the word shown in the VoiceList in FIG. 6 is searched from the synonym dictionary 252, and those that match the Category are extracted and stored as candidate words. The recognition processing unit 251 executes speech recognition processing of the input speech from the user using the speech recognition candidate words given from the candidate word extraction / storage unit 253. A flowchart example of candidate word extraction / storage processing is shown in FIG.
[0033]
The browser device 220 edits the processing result of the automatic input item in the automatic input processing unit 229 and the processing result of the user input item in the voice recognition unit 250 and returns a response to the center facility 100 through the communication unit 210 (that is, transmits the response). ).
[0034]
The configuration and operation of one embodiment of the present invention have been described in detail above, but the present invention is not limited to this, and various modifications and expansions are possible.
[0035]
【The invention's effect】
As described above, according to the present invention, the following effects can be obtained.
(1) By appropriately defining and using a content description language and its browser, it becomes possible to transmit useful information from the in-vehicle environment while ensuring safety.
(2) By converting the content description format according to the terminal type, compatibility with conventional HTML-based content can be maintained, and access from the office / home can be performed in exactly the same manner as before.
(3) The content description language according to the present invention incorporates the features of the conventional HTML, and the content writer newly learns with a format and the like, so that the content description method according to the present invention is easy to spread. become.
(4) When the system of the present invention is applied to, for example, a bulletin board, the bulletin board relates to a specific category, and the contents / format described therein are relatively standardized. Therefore, most information transmission contents can be covered by the key extraction according to the present invention.
(5) With the recording function on the server side, speech recognition of free sentences can be performed on the server side that enables large-capacity dictionaries and large amounts of processing, recorded sound can be distributed directly as content, In addition, surrounding sounds (festival sounds, sea sounds, animal calls, etc.) themselves can be delivered as content.
(6) Candidate words for voice recognition can be efficiently limited, and the voice recognition rate can be improved even in an in-vehicle environment that is a severe environment for voice recognition such as road noise.
(7) Parameters such as position, time, and speed can be automatically input, and information can be transmitted even during driving without bothering the user.
(8) It is possible to collect information from the center side without bothering the user and share it in traffic jam information or sightseeing information.
(9) By collecting the information of the terminal browser by utilizing the function of User Agent, it is possible to distribute the contents according to the correspondence of the terminal browser to the method of the present invention and enjoy the maximum possible benefit for the user become.
[Brief description of the drawings]
FIG. 1 is a block diagram of a system configuration according to an embodiment of the present invention.
FIG. 2 is a diagram showing an example of content database registration according to the present invention.
FIG. 3 is a diagram illustrating a characteristic example of an in-vehicle environment.
FIG. 4 is a diagram showing an example of content transmitted from the center facility to the vehicle terminal facility.
FIG. 5 is a diagram illustrating a classification example in the decoder of the browser device for the content of FIG. 4;
FIG. 6 is a diagram illustrating a process description example in a calculation unit of the browser device.
FIG. 7 is a diagram showing an example of a flowchart of candidate word extraction / storage of a speech recognition unit.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 100 Center equipment 110 Transmission processing part 120 Item extraction part 130 Content database 140 Acoustic processing part 150 Communication part 200 Vehicle terminal equipment 210 Communication part 220 Browser apparatus 230 Calculation part 240 Various sensors 250 Voice recognition part 260 Voice synthesis part 270 Various user interfaces

Claims (7)

An information exchange method in which an in-vehicle terminal arbitrarily communicates with an information storage / providing center including a content database to exchange information,
The center side determines the type of the in-vehicle terminal of the communication partner from the information of the User Agent in which the type of the in-vehicle terminal is described, and transmits the content of the question text in the description format according to the type to the in-vehicle terminal.
The in-vehicle terminal side analyzes the content of the question sentence sent from the center side, automatically transmits sensor information to the center side according to the contents of the input items of the content, or receives information input from the user Outgoing,
The center side processes information transmitted from the in-vehicle terminal for the content of the question sentence, and registers and accumulates the result in the content database.
An information exchange method characterized by that. The information exchange method according to claim 1,
The description format of the content of the question sentence transmitted from the center to the in-vehicle terminal is upward compatible with HTML, and is a description format capable of controlling the in-vehicle device such as automatic input of sensor information, speech synthesis and recognition An information exchange method characterized by In the information exchange method according to claim 1 or 2,
The in-vehicle terminal side analyzes the content of the question text sent from the center side, and when the items that can be acquired by the in-vehicle sensor such as position, time, speed, illuminance are instructed in the content, An information exchange method characterized by transmitting the value of an item acquired from an in-vehicle sensor without necessarily confirming the user. In the information exchange method of Claims 1 thru | or 3,
The in-vehicle terminal side analyzes the content of the question text sent from the center side, uses speech synthesis and speech recognition, and from the contents described in the content in the input items by conversation with the user, An information exchanging method characterized in that candidate words are dynamically constructed and used for speech recognition. The information exchange method according to any one of claims 1 to 4,
The in-vehicle terminal side sends the voice as it is to the center side in the free text input item of the question text content,
The center side performs voice recognition of the voice transmitted from the in-vehicle terminal side, registers the recognition result in the content database, accumulates the voice separately, and links to the recognition result of the content database. method of exchange. The information exchange method according to any one of claims 1 to 5,
On the center side, free text data input from the office / home terminal is analyzed to extract words and phrases that can be the contents of the question sentence content items, and words and phrases that can be expired from the data. An information exchanging method characterized by registering data including an expiration date in a content database. An information exchange system in which an in-vehicle terminal communicates arbitrarily with an information storage / providing center to exchange information,
An information exchange system, wherein the in-vehicle terminal and the information storage / providing center have a function of realizing the method according to any one of claims 1 to 6.

Images