KR20020072359A - System and Method of manless automatic telephone switching and web-mailing using speech recognition - Google Patents

Abstract

The present invention relates to an unattended automatic telephone exchange and webmailing system and method using voice recognition, comprising: an IVR module comprising an automatic voice recognition (ASR) middleware, a voice synthesis (TTS) middleware, and a call controller (interactive voice) Response module), a message control module that converts a caller's input voice message into a webmail using MIME and transmits it to a specific mail server, and a user information storage module that stores information of a user of a communication terminal connected to a private exchange. Is made of.

By using the system and method according to the present invention, it is possible to exchange information with the called party simply by inputting the information of the called party by voice. In addition, by providing the absentee webmailing function, the huge voice message storage means necessary for the existing absentee voicemail function is not required, and the callee can use the existing caller's own mail system without access to a separate exchange system. It is convenient because you can check voice messages.

Description

System and Method of manless automatic telephone switching and web-mailing using speech recognition}

The present invention relates to an unmanned automatic telephone exchange and mailing system using voice recognition, and a method of using the same, more particularly, to automatically exchange a telephone by recognizing a caller's voice, and in the absence of a telephone receiver. The present invention relates to a system and a method for transmitting voice or text mail to a user.

The current Private Branch eXchange (PBX) is a telephone exchange that allows all employees to share a certain number of external telephone lines used in companies, and to automatically connect telephones among internal users connected to the extension. System. The primary purpose of using PBX is to reduce the cost of assigning a regular telephone line to all employees in the company. Originally, analog technology was used, but nowadays digital technology is applied. A private branch exchange (PBX) is a telephone trunk line terminated in a PBX, a computer used to exchange and manage calls inside and outside the PBX (requires its own memory), a circuit network within the PBX, and a console or switchboard for the operator. It is manufactured and sold by domestic companies such as Samsung Electronics, LG Information and Communication, and Lucent Technologies, Northern Telecom (NORTEL), Rolm / Siemens, NEC, GTE, Intecom, Fujitsu, Hitachi, Mitel.

When using such a private telephone exchange, the operator makes a mechanical call to the called party's phone, or when the caller calls the telephone number using the Audio Response System (ARS), the PBX uses the ARS. The announcement is sent to the caller, and the caller is used in the manner of connecting with the desired caller by entering the desired CO line number.

However, according to a conventional method using an ARS or using an ARS, the caller must know the extension number of the caller, and it is difficult to find out the extension number corresponding to the caller's person or department name.

In addition, when connected to the desired caller, if the caller is absent, he cannot make a call or leave a message to the employees nearby. Recently, some private exchanges have a voicemail function. When there is no caller, the caller records a voice message to the caller and stores it in his / her mailbox. There are also services that allow you to listen to messages left behind. However, even in this case, the called party can be troublesome because it is possible to check whether or not the message came to him only when he is connected to the private exchange by telephone. In addition, since a large number of voice messages need to be stored, a huge storage space is required, which incurs high costs.

Therefore, the present invention was conceived to overcome these disadvantages, by using speech recognition (Speech Recognition) can be easily connected to the desired caller only by voice, the voice message of the caller to the webmail server that the caller uses the existing web Provides a system that can be sent in the form of mail.

It is an object of the present invention to provide an unmanned telephone automatic switching system using a voice recognition system and a method of using the same.

Another object of the present invention is to provide a voice recognition auto exchange / webmailing system and method that enables a caller to be connected to the caller by simply inputting the information of the caller by voice.

It is another object of the present invention to provide a method for transmitting a voice message of a caller in real time to an Internet mail server used by the caller when the caller is absent, either automatically or at the option of the caller.

Figure 1 shows the configuration of the entire system for implementing the voice recognition automatic exchange / web mailing method according to the present invention.

Figure 2 shows the detailed configuration of the voice recognition automatic switching / web mailing system according to the present invention for each module.

Figure 3 shows the overall flow of the voice recognition automatic exchange / web mailing method according to the present invention.

The system for achieving the object of the present invention as described above has the following configuration.

A system connected to a mail server through an internet network and a private exchange (PBX) through an extension system. An IVR module comprising an automatic speech recognition (ASR) middleware, a speech synthesis (TTS) middleware, and a call controller ( an interactive Voive Response Module), a message control module that converts a caller's input voice message into a webmail using MIME and transmits it to a specific mail server, and user information that stores user information of a communication terminal connected to a private exchange. It consists of a storage module.

The above-described message control module is activated when the desired talker is absent, and converts the original voice input by the talker into a wav file and converts it into a file that satisfies the MIME standard. And a mail server search means for searching the user information storage module to search the webmail server of the corresponding receiver, and a mail transmission means for transmitting the converted mail file to the mail server. Each means constituting the message control module may be implemented in hardware or software.

In addition, the voice recognition auto-exchange / webmailing method according to the present invention uses the above-described system and consists of the following steps.

Determining, by the call controller, whether or not the callee's search input is in a voice form or a DTMF format, and recognizing a voice input using a voice recognition middleware only when the input is a voice, and searching for a user information storage module. Extracting the extension number of the party, determining whether the caller is absent, receiving a voice message to be delivered from the talker, and converting the voice message into a webmail form using a message controller; The controller searches for the user information storage module and transmits the voice mail message in the form of the web mail via the Internet to the extracted mail server of the called party.

Hereinafter, with reference to the accompanying drawings will be described in detail an embodiment of a voice recognition automatic exchange / web mailing system and method according to the present invention.

1 shows the configuration of an entire system for implementing the method according to the invention.

The entire system includes a public switched telephone network (PSTN) 13, a private telephone exchange (PBX) 14, a voice recognition auto attendant / webmailing system 10, the internet 16, and the web. It consists of a mail server 17. In addition, a plurality of communication terminals 18 are connected to the private exchange, and the private exchange is connected to the external wired / wireless communication terminals 11 and 12 through the PSTN.

The talker uses external wired / wireless communication terminals 11 and 12 and connects to the private exchange via the PSTN network. The talker listens to the announcement by an ARS (Automatic Response System) service provided by a private exchange or the system according to the present invention, and inputs the information of the caller who wants to talk by DTMF (Dual Tone Multi Frequency) signal or voice using a button. .

In the case of inputting the DTMF signal, the private exchange may switch to the called party's terminal according to the conventional method. In the case of voice input, a service provided by the system according to the present invention (described in detail with reference to FIG. 2) is received.

The wired / wireless communication terminals 11 and 12 may be devices generally used for voice communication such as general wired telephone, PCS (personal communication system), cellular phone, and IMT-2000, but in particular, the Internet network and VoIP (Voice) It may also be a computer system such as a PC, a portable notebook, a personal digital assistant (PDA), a handheld PC, etc., which enables voice data communication using an over IP solution. In addition, although not shown, a mobile communication system such as a base station, a repeater, and a switch interworking with a PSTN network may be further provided for wireless communication.

The private exchange (PBX) is connected to the PSTN network through a trunk line (indicated by a double line), and provides a service according to the present invention and a communication terminal (mainly a general fixed line telephone) of various users through an extension (indicated by a single solid line) It is linked with the system 10.

The voice recognition automatic exchange / webmailing system 10 according to the present invention is connected to a plurality of webmail servers 17 via the Internet 16. That is, data communication between the system 10 and the webmail server 17 according to the present invention is performed through a computer network connection channel 18 such as the Internet, and uses a TCP / IP based protocol. The protocol may use TCP / IP, User Datagram Protocol (UDP) as it is, or may use a separate protocol implemented according to the present invention through an application level disposed in a higher layer of TCP / IP.

The Internet has many services in the TCP / IP protocol and its upper layers: HyperText Transfer Protocol (HTTP), Telnet, File Transfer Protocol (FTP), Domain Name System (DNS), Simple Mail Transfer Protocol (SMTP), and SNMP ( Although it refers to a global open computer network structure that provides Simple Network Management Protocol (NFS), Network File Service (NFS), and Network Information Service (NIS), in the present invention, SMTP of MIME (Multipurpose Internet Message Extension) standard is mainly used. .

The mail server 17 is a plurality of general servers implemented by the prior art, and means a system for relaying mail exchanges between users.

Since the above-described PSTN network, private exchange (PBX), mail server 17, etc. are sufficiently supported by the prior art, detailed description thereof will be omitted.

Figure 2 shows the detailed configuration of the voice recognition automatic exchange / web mailing system according to the present invention for each module.

The system according to the present invention largely comprises an IVR module 20, a message control module 25, and a user information storage module 26. The IVR module 20 includes a call controller 21 connected to a private exchange (PBX), a message control module 25, and a user information storage module, a text-to-speech (TTS) middleware 22, and an ASR (Automatic). Speech Recognition) middleware (23). The message control module 25 is linked with a plurality of mail servers through a TCP / IP based internet network.

Although not shown, the message control module 25 again converts the webmail means, the mail server search means for searching the user information storage module to search the webmail server of the corresponding receiver, and converts the converted mail file into the corresponding mail server. And a mail transmission means for sending the message.

Each module may be implemented in the form of hardware, software, and a combination thereof. The hardware used in this embodiment uses an operating system (OS) of Windows NT 4.0 and Window 2000, and uses an Intel Pentium III-500MHz central computing unit. It is a computer system with 256MB main memory, 10GB auxiliary memory, D / 41ESC voice board and analog telephone line.

The IVR module and the message control module are implemented in software by Visual C ⁺⁺ , using L &H's ASR1500 solution for speech recognition, L &H's RealSpeak V2 solution for speech synthesis, and 8-bit supporting ADPCM and PCM. A-law, μ-law of the voice format was used.

The user information storage module 26 includes general information about a user, a department, or a communication terminal managed by a private exchange, and is implemented in hardware or software. For example, it may be in the form of a database (DB) that is implemented in the auxiliary storage device in an appropriate data structure to be managed by a specific database management system.

When the user information storage module 26 is implemented as a database, a relational database management system (RDBMS) such as Oracle, Infomix, Sybase, DB2, Gemston, An object-oriented database management system (OODBMS) such as Orion, O2, etc. may be implemented for the purposes of the present invention, and includes a user name, affiliation department, extension number, and web mail account (mail address). Must have appropriate fields. In this embodiment, the user information storage module is implemented as a database using MS SQL (Microsoft SQL 2000).

The following describes the function of each module in detail.

The call controller 21 determines whether the signal input by the talker is in the form of voice or DTMF, and in the case of the DTMF signal, connects to the extension of the called party through a private exchange (PBX). The ASR middleware is operated to recognize the voice, search the user information storage module, and check whether there is a user that matches the input voice signal. When the call controller searches for the extension number of the corresponding user, the call controller forms a communication channel between the communication terminal and the caller of the corresponding user by driving the private exchange. At this time, the search may be performed, for example, by recognizing a voice, converting it into a digital signal in the form of text or ASCII code, and searching for user information stored in a uniform format.

In addition, when the call controller 21 determines that the connected receiver is absent due to not receiving a call for a predetermined time, the call controller 21 drives the TTS middleware (or ARS system) to output a guide voice to the caller. For example, check whether you want to deliver the message using webmail. When the talker wants to send a webmail message, the talker receives a voice message from the talker and transmits the message signal to the message control module 25.

The message control module 25 converts the voice message signal received from the call controller into a WAV file and encodes the file into a file compliant with the MIME standard to form a webmail file. WAV file conversion and MIME type webmail file generation are described in detail below.

The message control module 25 forms a webmail file, retrieves the user information storage module 26, extracts the mail address of the corresponding talker who is absent, and sends the webmail file to the corresponding mail server via the Internet network. send. At this time, the call controller has already searched for the user information storage module, that is, the user name or extension information can be transmitted to the message control module along with the voice message to be used as the search data of the message control module. The call controller can also directly extract the mail address and forward it along with the voice message to the message control module. In such a case, the message control module does not need to search the user information storage module and transmits the converted webmail to the corresponding mail address.

Although not shown or described in detail, the systems and methods of the present invention may provide other services utilized by conventional voice recognition auto-exchange methods. For example, you can connect to another communication terminal that you designate while you are away, a voice mail function that stores the message of the caller when you are absent, and tells the caller when you are away. Out of office greetings and other user preferences.

The method of converting the original audio message into the WAV format is based on the prior art.

First, to convert analog voice data directly, wav file, MIDI file, mixer program, wav file editor program, and sound card are used. But,

In order to form a higher quality wav file, it is desirable to use a digital sound card that converts analog sound into digital through an A / D converter or DAT in advance, and inputs the converted digital signal and stores it as a wav file. Do. Digital sound cards for this purpose include Card Plus, Digital Only Card, and ZA2 cards from Zeifiro.

MIME is a new standard designed to improve the traditional method of sending plain e-mail textual e-mail only. MIME allows you to convert binary files such as program, video, voice, and graphics into text-only mail messages. Can be sent to. Most modern web browsers are aware of MIME content and can easily be configured to support additional MIME types. Therefore, when a mail program or a web browser receives a MIME message including, for example, an MPEG video clip, a sound format of WAV format, the original MPEG data and WAV data can be reconstructed from the text of the mail message.

MIME defines additional fields in addition to Internet mail headers. Additional fields include the MIME-Version field, which indicates the MIME version, the Content-Type field, which indicates the MIME type, and how to convert data of a particular type into ASCII text. A Content-Transfer-Encoding field, a Content-ID field for assigning a unique ID, a Content-Description field as a description part of MIME data, and a Content-Type field subdivided into types and subtypes.

Content-Type field related to voice uses "audio" as type, "basic" for audio data stored in 8-bit ISDN mu-law format as subtype, and "midi" for audio data stored in MIDI format. &Quot; wav " is used for audio data stored in WAV format.

In the webmail conversion according to the preferred embodiment of the present invention, since the digital or analog raw voice message is converted into a WAV format and converted into a webmail file conforming to the MIME standard, "audio" "wav Use the Content-Type field of ".

In addition, a "base64" or "quoted-printable" encoding method may be used to convert a WAV format voice message to ASCII text format, but is not limited thereto.

Speech Recognition technology used in the present invention briefly describes the process to the extent that it can recognize hundreds or thousands of commands (search words) input by an unspecified speaker. First, when the voice input through the microphone is converted into a digital signal and enters the voice recognition system, the voice feature that can express the phonetic characteristics every 10-20 msec through the voice feature analysis process, which is the first step of the voice recognition The coefficients are extracted. The extracted speech feature coefficients are passed to the pattern recognition process and compared with the models of pre-stored words or phonemes. To do this, first, basic units of words (words, semi-syllables, phonemes, and transition sounds) for pattern recognition are determined. However, it is necessary to obtain and store each representative pattern or model corresponding to these speech units in advance from the training speech data. Next, when the feature pattern of the input voice to be recognized is analyzed, the speech units corresponding to the closest patterns are determined as candidates of the recognized words or phonemes by comparing the feature patterns with the stored representative patterns or models. There are many ways to recognize this pattern, but a statistical method called Hidden Markov Model (HMM) is widely used. This voice recognition technology is similar to that used in the voice dialing technology of the conventional mobile phone, the present invention used the ASR1500 solution of L & H as described above.

Speech Synthesis refers to a process of converting a sentence input in text form into a speech signal. In accordance with the text stored in the form of keyboard input, a character recognition device, and a database pre-stored in a computer, a method of reproducing a combination of voice signals corresponding to each pre-stored word, phrase, or sentence. In the embodiment, L & H's RealSpeak V2 solution was used for speech synthesis. Since the speech recognition and speech synthesis can be implemented by conventional techniques, the detailed description thereof will be omitted.

Figure 3 shows the overall flow of the voice recognition automatic exchange / web mailing method according to the present invention.

When the caller connects to the call controller via the PSTN network and the private switch, the call controller transmits a guided voice that allows the user to select whether to enter the extension number with a button or to exchange it by voice. The call controller 21 determines whether the input exchange information is voice (S31), and in the case of the voice signal, recognizes the voice using an ASR middleware and converts the exchange information into other appropriate data. The IVR module searches for the user information storage module (S32) and checks whether there is the same user as the converted exchange information data (S33), and if so, drives the private exchange to connect to the corresponding extension number (S37).

If the caller does not exist, the IVR module synthesizes the guide voice that the requested caller does not exist using the TTS middleware or other ARS system, and outputs it to the caller (S40, S41).

If the caller terminal is not connected to the callee terminal for a predetermined number of times or a predetermined number of times, the IVR module determines that the caller is absent (S35) and transmits a guide voice asking whether to leave a voice message to the caller. When the caller inputs a voice message, the IVR module receives the voice message and delivers the voice message to the message control module, and the message control module converts the voice message into a wav type MIME webmail (S36).

The message control module retrieves the user information storage module directly, or obtains the webmail address of the callee received from the IVR module, and transmits the previously generated MIME type voice message webmail to the corresponding mail server via the Internet. (S38, S39).

By using the automatic exchange / webmailing system using voice recognition according to the present invention, the caller can store the extension number by calling the caller simply by inputting the name or department name of the caller. You do not have to do.

In addition, by providing an absentee webmailing function, which is not a function of the conventional private exchange, a huge voice message storing means necessary for the existing absentee voicemail function is eliminated.

In addition, it is convenient for the employee (receiver) to check the voice message by using the existing mail system that he always uses in the Internet environment without accessing the voice mailbox and checking the message.

Claims (3)

It is a system that is connected with mail server and private exchange (PBX) through internet network. An IVR module comprising an automatic voice recognition (ASR) middleware, a voice synthesis (TTS) middleware, and a call controller; A message control module for converting an input voice message of a talker into a web mail form and transmitting the same to a specific mail server; Unattended automatic telephone exchange and webmailing system using a voice recognition, characterized in that consisting of; employee information storage module for storing information of the user of the communication terminal connected to the private exchange. The method of claim 1, The message control module is activated when the desired receiver is absent, Webmail converting means for converting the original voice input by the talker to wav file, and converts the original voice into a file that satisfies the MIME standard to generate a webmail, A mail server search means for searching a user information storage module and searching a webmail server of a corresponding receiver; Unattended auto attendant and webmailing system using voice recognition, characterized in that it comprises a mail transmission means for transmitting the converted mail file to the mail server. IVO module which is connected to PBX through extension and mail server via internet, and consists of automatic voice recognition (ASR) middleware, voice synthesis (TTS) middleware, and call controller And a system comprising a message control module and a user information storage module, Determining, by the call controller, whether a caller's callee search input is in a voice form or a DTMF format; Recognizing a voice input using a voice recognition middleware only when the input is a voice, and searching for a user information storage module to extract the extension number of the called party; Determining whether the called party is absent, if not, receiving a voice message to be delivered from the talker, and converting the input voice message into a webmail form using the message controller; And the message controller searching for a user information storage module and transmitting the voice message of the webmail form to the extracted mail server of the corresponding receiver via the Internet. Webmailing method.