KR100660293B1 - Terminal voice menu moving system - Google Patents

Abstract

Disclosed is a terminal voice menu mobile system capable of providing various contents to users by adding a voice recognition-based menu navigation function to an application using a mobile communication terminal. The system includes a mobile communication terminal for encoding and outputting a sound source uttered by a user; A back-end unit which receives the encoded sound source and interprets it, and then outputs a command signal according thereto; And a content server providing data desired by a user according to the command signal.

  Terminal Applications, Speech Recognition, Voice Menu Navigation, Distributed Speech Recognition (DSR), Automatic Speech Recognition (ASR)

Description

Terminal voice menu moving system {SYSTEM FOR NAVIGATING A VOICE MENU IN A MOBILE PHONE AND METHOD THEREOF}

1 is a view showing the overall configuration of a mobile terminal voice menu moving system according to the present invention.

FIG. 2 is a diagram for explaining a function of a DSR server in FIG. 1.

FIG. 3 is a diagram for explaining a function of an ASR server in FIG. 1.

4 is a view for explaining the function of the WA server.

FIG. 5 is a diagram for explaining a function of a DSR API in FIG. 1.

6 to 8 are display screens showing a menu moving process according to the present invention.

8 is a view showing a display screen of the N-best selection function according to the present invention.

9 is a view showing a use example of the terminal voice menu moving system according to the present invention.

<Description of Symbols for Main Drawings>

10: mobile communication terminal 11: DSR API

20: back-end part 21: DSR server

22: ASR Server 23: TTS Server

24: VXML Interpreter 30: Content Server

The present invention relates to an additional service of a mobile communication terminal, and more particularly, to a terminal voice menu mobile system for providing various contents to users by adding a voice recognition-based menu navigation function to an application using the mobile communication terminal. It is about.

Recently, with the development of mobile communication technology, voice recognition and voice processing technology has been activated for voice recognition dialing service to make a call by inputting the caller's name by voice without directly dialing the telephone number.

In addition, the need for additional services using voice recognition to obtain information is increasing day by day.

On the other hand, the operation of the mobile communication terminal can be roughly divided into simple operation and menu operation of various services. Conventional mobile communication terminals have services such as voice recognition dial service, voice recognition content service, phone information upload service, and phone information download service. In case of using, the user accesses the service by inputting the service access number (for example, 15 **-****) or the shortcut number (# **) by using a combination of two or more digits. .

However, in this case, the user has to remember the number and then has to enter the number one by one.

As another example, in the case of a road guidance service employing the existing voice recognition technology, a user operates a button to access a corresponding service, selects a desired menu, and then speaks a destination by voice. In this case, since many operations must be done manually by the user until the use of the service, use of it during driving may cause a lot of inconveniences and risks.

Even after accessing the service, there was an inconvenience in that the user must find a corresponding menu according to the usage method and operate a key. When using the voice recognition dial service, the user accessed the personal address book of the subscriber database in the web environment, and registered, edited, and deleted the phone number. However, this could not be managed in a place where the Internet environment was not provided.

Accordingly, there is a demand for a method of overcoming such an unreasonable point of use of the additional service and enabling the user to use the additional service more simply and efficiently.

The present invention has been made to solve the problems according to the prior art as described above, an object of the present invention is to add a voice recognition-based menu navigation function to the application using a mobile communication terminal without various construction restrictions to users The present invention provides a terminal voice menu mobile system that can provide contents.                         

It is another object of the present invention to apply a voice recognition technology to a mobile communication terminal so that a user can receive all the key operations and various additional services of the mobile communication terminal using his or her own voice instead of manipulating buttons one by one. It is to provide a voice menu moving system.

In order to achieve the above object, a terminal voice menu moving system according to the present invention comprises: a mobile communication terminal for encoding and outputting a sound source uttered by a user; A back-end unit which receives the encoded sound source and interprets it, and then outputs a command signal according thereto; And a content server providing data desired by a user according to the command signal.

Before describing an embodiment of the present invention with reference to the accompanying drawings, it will be described the internal environment of the mobile communication terminal to be assumed in order to implement the terminal voice menu mobile system according to the present invention as a first priority embodiment.

First, for speech menu navigation, a DSR (Distributed Speech Recognition) Protomming Interface (DSR API) must be built and the DSR API can be easily used, and a speech recognition engine is built into the terminal for initial menu recognition. It is important to establish a system that can add voice recognition menu navigation to all applications. In addition, the current voice platform, which is only isolated language recognition, needs to be upgraded to natural language recognition, and a grammar that can be directly applied to services by tuning the recognition engine should be constructed.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a view showing the overall configuration of a mobile terminal voice menu moving system according to the present invention.

As shown in the drawing, in the terminal voice menu moving system according to the present invention, the Distributed Speech Recognition (DSR) Application Progromming Interface (API) 11 of the mobile communication terminal 10 includes a back-end unit 20. Communication with the DSR server 21, the DSR server 21 is an Automatic Speech Recognition (ASR) server 22, a Text-To-Speech (TTS) server 23, and Voice eXtensible Markup Language (VXML). The voice extensibility notation language) supports the service from the contents server 30 in conjunction with the interpreter server 24.

Each device of the voice menu moving system according to the present invention configured as described above will be described with reference to FIGS. 2 to 5.

FIG. 2 is a diagram for explaining a function of a DSR server in FIG. 1.

As shown therein, the DSR server 21 simultaneously performs voice output / screen output through interworking of VoiceXML and the terminal application.

Meanwhile, the DSR server 21 supports a DSR-based Multimodal UI capable of simultaneously performing voice input / key input on the mobile communication terminal 10 using a data channel. In addition, organic communication with the mobile communication terminal 10 is possible. That is, the DSR server 21 is capable of high performance of the voice engine, and by using the voice call codec (EVRC), tuning is easy without changing the hardware of the mobile communication terminal 10.

FIG. 3 is a diagram for explaining a function of an ASR server in FIG. 1.

The ASR server 22 implements and operates a service of a voice recognition engine provided in a library form. The ASR server 22 is implemented in the form of a server on TCP, and applications that require voice recognition services become clients and use the services of the ASR server 22.

In addition, the ASR server 22 uses one or more TCP ports to provide a service to a client, and allows a plurality of TCP socket connections to allow a plurality of clients to use the service at the same time.

The ASR server 22 provides a voice recognition service through a virtual voice recognition channel. A channel is a unit that uses the service of the ASR server 22 separately from the TCP socket, and is distinguished by a unique ID. The ASR server 22 supports the use of multiple channels with one TCP socket, and the maximum number of available channels is limited by the ASR server 22.

The ASR server 22 stores the result data in its server before completing the requested voice recognition task and sending the result to the client. This stored data is used for continuous tuning of the speech recognition engine.

The ASR server 22 is executed through a console window, and the executed program is continued without being terminated in a normal case. At this time, you can check the current channel status, I / O and voice recognition results through the executed console window.

FIG. 4 is a diagram illustrating a function of a WA server. The WA server refers to a server of a middleware concept that enables a CP to remotely upgrade a grammar required for a service to a mobile communication terminal. And, since it is already known in the art, a detailed description thereof will be omitted.

FIG. 5 is a diagram for explaining the function of the DSR API in FIG. 1, and the DSR API 11 performs the following functions.

(1) create a voice recording

When a signal is received in response to a recording request, the recording source API is recorded using a recording related API of the Wireless Internet Platform for Interoperability (WIPI). The recorded sound source is encoded by EVRC by the DSP (Digital Signal Processor) of the terminal and then stored.

(2) Voice start point endpoint detection

The user presses the button once at the beginning of the talk, and continues recording even after the button is released. Recording stops automatically when the user does not speak for more than a predetermined time. Based on the Zero Crossing Rate and dB, the captured sound source detects a start point and an end point.

(3) voice level display

When the voice recording is in progress, a small icon indicating the volume of the voice recorded at a specific position on the terminal screen is displayed.

(4) Create N-best screen

As a result of the voice recognition by the DSR server 21, when multiple answers are given, the candidates are properly displayed on the screen. On this screen, the user can simply select a number or select it or directly press the Dual Tone Multi Frequency (DTMF) button.

(5) TCP communication with DSR server

The sound source encoded by EVRC is transmitted to the DSR server 21 through a TCP connection. The DSR server 21 also drops the voice recognition result through this TCP connection.

Next, a process of moving a menu according to food recognition using the apparatus of each unit configured as described above will be described.

6 to 8 are display screens showing a menu moving process according to the present invention, FIG. 6 is a basic interface process according to voice recognition, FIG. 7 is a voice command system from a user, and FIG. 8 is an N-best selection. Each screen is a display screen.

First, referring to FIG. 6, a basic interface process according to voice recognition will be described.

The application of the terminal displays a lip icon on the screen when the voice recognition is possible. When the user presses the voice input button ("call" button), the volume of the voice being recorded is expressed in the form of an icon at the bottom of the screen.

While voice recording is completed and the recorded voice is handed over to the DSR server 21 for recognition, the user can grasp the progress of voice recognition by placing an icon of "voice recognition" at the bottom of the screen. do. Voice recording is initiated with a confirmation tone when pressed to turn on from off with a separate recognition hotkey. At this time, a progress bar displaying the remaining voice recording time is displayed for about 2 or 3 seconds and the end point detection is also performed.

If the recognition process is successful, the menu is highlighted to allow the user to feed back the recognition result. If the recognition result is uncertain, a small window is displayed at the bottom of the screen to process N-best. Have the user make a final check.

Next, a menu item moving process according to a voice command system from a user will be described with reference to FIG. 7.

First, a voice command is divided into a simple voice command for reading a currently displayed menu item and a short voice command crossing a menu tree of several levels.

Simple voice commands are made by the user reading a menu item on the screen. In the case of a menu item composed of compound words, various alternative labels should be considered. For example, a menu such as "Find Peripheral Facility" may be abbreviated in one word by the user, such as "Peripheral Facility", so that various alternative labels (Alias) are added to Grammar for the user's convenience.

The short voice command is a function of collecting a utterance pattern (Corpus) of a user who uses an application of the terminal and giving a command at a time without going through a menu step. For example, for voices such as "Directions to Seoul City Hall", a user can be selected by selecting a "Route Guidance" and a POI called "Seoul City Hall".

In order to support the short voice command, the speech recognition engine server must be able to recognize the continuous word, and the grammar must be able to accommodate the grammatical structure of the Augmented Backus-Naur Form (ABNF) form or the equivalent form.

On the other hand, the voice recognition of the DSR structure has a problem that the user's waiting time may increase due to the limitation of the transmission speed of the wireless Internet while going through a complicated process of encoding, transmitting, decoding, recognizing, and transmitting a recognition result. In general, the process is completed within 1 to 2 seconds, but the first connection connection process is delayed.

Since the voice recognition time delay may be a major cause for the user to evade the voice recognition application of the terminal, at least the initial screen may be recognized using the voice recognition engine built in the terminal to minimize the user's waiting time. Apply the technique.

The speech recognition engine embedded in the terminal has a problem in that a large vocabulary grammar or continuous word recognition is not supported due to the limitation of the terminal specification. Therefore, in the case of the initial screen recognized by the terminal, only the menu items displayed on the screen and some alternative labels are included in the speech recognition vocabulary.

Some users may prefer to skip menus via short voice commands, even if there is some initial delay. Accordingly, the initial screen may be provided with a setting user interface for determining whether to use a voice recognition engine built in the terminal or to use a more sophisticated and rich voice recognition function through the DSR server.

8 is a view showing a display screen of the N-best selection function according to the present invention.

Since the speech recognition of the system cannot satisfy 100% recognition rate, it must provide a friendly user interface for out-of-vocabulary.

If a word spoken by the user does not show an overwhelming accuracy of any of the words to be recognized, a function of narrowing the object to a list of words having a relatively high accuracy (60%) is required to guide the user to speak again. This function is defined as N-best Selection function.

When N-best is selected, the number of target recognition words is reduced so that the user can predict a higher recognition rate even if the user has the same voice.When the N-best screen is displayed, the user presses a button to enter or “1”, You can expect accurate recognition because it is more likely to try to input differently by uttering a number like "2".

In addition, through the N-best selection function, the long timeout for the recognition error is not set, and the determination of the recognition error is made as soon as possible, thereby reducing the waiting time of the user.

9 is a view showing a use example of the terminal voice menu moving system according to the present invention.

As shown, the user can press the call button once and see that all services are spoken. As such, when the content is used by voice, unnecessary operation is eliminated, and the user can use the service comfortably and safely.

While the invention has been shown and described with respect to certain preferred embodiments thereof, the invention is not limited to the embodiments described above, but in the field to which the invention pertains without departing from the spirit of the invention as claimed in the claims. Any person with ordinary knowledge will be able to make various modifications.

The greatest advantage of the present invention is that the user makes a voice selection of all functions of the mobile phone and menus of various contents with a single button operation. This means that users can use various mobile phone services while doing other tasks.

In addition, according to the present invention, the user can see that all the services are used as voice after pressing a call button once. Therefore, when the content is used by voice, unnecessary operations are eliminated, and the user can use various additional services provided by the communication company conveniently and safely.

Claims (9)

A mobile communication terminal for encoding a sound source uttered by a user with an enhanced variable rate codec (EVRC) and outputting the encoded sound source; A back-end unit which receives the encoded sound source and interprets it, and then outputs a command signal according thereto; And And a content server providing data desired by a user according to the command signal. The method of claim 1, wherein the mobile communication terminal, Implemented as a Distributed Speech Recognition (DSR) application programming interface (API) for moving a voice menu, and a terminal voice menu moving system comprising a built-in voice recognition engine for initial menu recognition. The method of claim 2, wherein the DSR API, A terminal voice menu moving system, which performs voice recording generation, voice start and end point detection, voice level display, N-best screen generation, and TCP communication with the DSR server. The method of claim 3, wherein the voice recording is generated, And recording the sound source according to the user's recording request signal, and encoding and storing the recorded sound source. The method of claim 3, wherein the voice start point and the end point detection, The terminal voice menu moving system, characterized in that performed based on zero crossing point and divide (dB). The method of claim 3, wherein the voice level display, And a predetermined icon representing the volume of the voice at a predetermined position on the screen during voice recording. According to claim 3, The N-best screen generation, And if the DSR server recognizes a plurality of answers, the candidate words are selectively displayed on the screen. According to claim 3, The N-best screen generation, The terminal voice menu moving system, which can be implemented by a user inputting only a predetermined number to select a candidate word or directly pressing a dual tone multi frequency (DTMF) button. The method of claim 1, wherein the back-end portion, A Distributed Speech Recognition (DSR) server that performs voice / screen output through interworking of VoiceXML and a terminal application program; And Implemented in the form of a library, the terminal voice menu moving system characterized in that it comprises an Automatic Speech Recognition (ASR) server for operating a service of the speech recognition engine.

Images