CN1457180A - Method for realizing interacting voice response in IP network - Google Patents

Abstract

The invention discloses the method for realizing interactive type voice response (IVR) in IP network based on MGCP protocol and VXML standard. The invention through MGCP protocol controls voice call of IVR device in IP network. Using vocoder makes the IVR device support real time voice in IP network, letting the flow control of the IVR device follow VXML standard. Thus, the IVR device can run on IP network independently. The method raises flexibility of the IVR system, lowers, cost for developing the system, and reduces the difficulties of development. It is easy to configure out IP IVR devices to meet various requirement.

Description

Realize the method for interactive voice response in the IP network

Technical field

The present invention relates to realize in the multi-media voice communication field method of interactive voice response (IVR), particularly relate to method based on interactive voice response (IVR) in MGCP agreement and the VXML standard realization IP network.

Background technology

Be that a flow process is explained executive system in interactive voice response (the Interactive Voice Response:IVR) system nature, it mainly explains and carries out the flow file that is generated by service creation environment, and, carry out alternately with the user by the Communication Control platform.The typical case uses as Fig. 1, wherein:

SSP (Service Switch Point): service exchange point, be responsible for finishing the work that speech channel continues in the intelligent network, receive SCP (service control point) and issue its action command, handle the speech channel that is continuing with ad hoc fashion, thereby different telephone services is provided.

SCP (Service Control Point): SCP, realize the service control function in the intelligent network, and use the NO.7 Signalling exchange between the SSP.SMS (Service Manage System): business management system manages SSP and SCP.

SCE (Service Create Environment): service creation environment, be responsible for creating new service logic.

On public service telephone network (Public Service Telephone Network:PSTN), there are wide variety of different types of IVR equipment at present, for example telephone bank, 168 information service centers etc.These equipment all are directly to receive on the PSTN by trunk interface.Owing to relate to complicated signaling process and trunk interface on the PSTN, cause it on special hardware, to develop; To make that the system that develops has a cost higher with the correlation of hardware, and the shortcoming that system flexibility is relatively poor generally also needs to develop again after having changed hardware; Owing to relate to complicated signaling process, for the developer who is unfamiliar with the PSTN network signal, development difficulty is also very big simultaneously, generally needs the developer of specialty to carry out corresponding development.And the flow process of present IVR system control adopts the language of definition voluntarily mostly, can't with other IVR system intercommunication, this has just caused the user must be familiar with the flow process control language of the IVR system that adopted, is not easy to the application of IVR system.

Traditional gateway is used for realizing service intercommunication between broadband network and the legacy network, and the calling that gateway needs to set up between PSTN and the Internet connects; Can carry out the conversion of media formats and the conversion of signaling: H.323 gateway is carried out in Internet one side, MGCP or Session Initiation Protocol, carries out isdn signaling, ISDN PRI or Signaling System Number 7 usually in the PSTN side; Because the conventional gateway function is too complicated, can only support less user usually, therefore for the needs of system extension, gateway must be separated.

The key of separating gateway is exactly that the media conversion function is separated with gateway control function.Make gateway only bear simple media conversion function, complicated gateway control function is then carried out by the independently controlled entity outside the gateway (Call Agent or Media Gateway Controller).Its feature is: will control intelligence and focus on network, in promptly a spot of Call Agent, the intelligent network in its thinking and the telecommunications network is similar.The MGCP agreement is exactly a kind of agreement that proposes according to the basic thought that gateway function separates.Gateway model after the separation as shown in Figure 2, the implication of each assembly among the figure is as follows:

SG: SGW (Signal Gateway), finish on the PSTN the signaling of signaling (for example: Signaling System Number 7, R2 signaling, isdn signaling) on the IP network (for example: BICC, conversion SIGTRAN).

MG: media gateway (Media Gateway), finish on the PSTN the voice of voice (for example: T1, E1) on the IP network (for example: conversion RTP/RTCP).

MGC: Media Gateway Controller (Media Gateway Controller), realize complicated gateway control function, common conduct is controlled entity independently.

MGCP: MGCP (Media Gateway Control Protocol), MGC message and the control flow by this protocol definition realized control to MG.

VoiceXML (VXML) is the extension standards of XML, and its main application is: by content and the information of sound and telephone interview Internet, in will using to IVR based on the exploitation of WEB and content transplanting.Structural model such as Fig. 3, the implication of each assembly among the figure is as follows:

Archive server (Document Server) is deposited the VoiceXML document, the flow process that the user needs by the customization of editor VoiceXML document.

VoiceXML interpreter (VoiceXML Interpreter) is mainly finished the explanation of VoiceXML document.

Carry out platform (Implementation Platform), mainly finish the execution of VoiceXML document, and be responsible for and the user carries out alternately.VXML is a plurality of mutual by definition in a file, has reduced the interactive operation of client/server; The application developer is freed from the details that bottom, platform are correlated with; User interactions code (VoiceXML) and service logic (CGI scripts) have been separated; Strengthened the portability of service at difference execution platform.For content supplier, instrument supplier and platform provider, VoiceXML is an all-purpose language.

Summary of the invention

The invention reside at the deficiency that exists in the background technology, a kind of method based on interactive voice response in MGCP agreement and the VXML standard realization IP network is provided,, improve its application flexibility to reduce the development cost and the development difficulty of system.

Solution of the present invention is:

Interactive voice response (IVR) system is inserted in the IP network, and described IP network adopts the voice call control of MGCP agreement to interactive voice response system;

In interactive responses (IVR) system, vocoder is set, the voice that receive are decoded and the voice that send are encoded;

In interactive responses (IVR) system, set up voice VXML document and VXML document management module, and making the operation flow of described VXML document customization and the control flow of interactive responses set up mapping relations, described VXML document management module is responsible for reading, the VXML document of explanation and executive system appointment;

When the user passes through the audio call interactive response system, Media Gateway Controller in the IP network is finished calling signaling reciprocal process and is made the calling party and the interactive response system connection, interactive response system is according to user's demand for services, by the XML document administration module VXML document of system's appointment is operated, with control IVR and user's reciprocal process.

According to technique scheme; Described interactive response system inserts IP network by Ethernet interface; Described vocoder is ACM decoder and ACM encoder; Described mapping relations be the flow process control of interactive responses require and the main syntactic element of VXML standard between mapping relations.

The present invention solves the signaling control that IVR uses on the IP network by the MGCP agreement, solves the flow process control that IVR uses by VXML, has realized that with regard to the interface that adopts complete standard the IVR on the IP network uses like this.Because all based on standard agreement, not only simplified the development process in the practical application greatly, also can with all standard device intercommunications at present.Adopt method of the present invention, can develop IVR equipment on computers, overcome the shortcoming of existing IVR system effectively and reduce development cost greatly, and make system have stronger flexibility based on software.

Description of drawings

Fig. 1 uses schematic diagram for the typical case of IVR system on PSTN;

Fig. 2 is the illustraton of model of conventional gateway function after separating;

Fig. 3 is the structural model figure of VXML;

Fig. 4 is a call control procedure schematic diagram of the present invention;

Fig. 5 is an IVR system logic block diagram of the present invention;

Fig. 6 is for adopting the schematic diagram of networking of the present invention.

Embodiment

Present embodiment inserts interactive voice response (IVR) system in the IP network by Ethernet interface, and IP network adopts the voice call control of MGCP agreement realization to interactive voice response system.The MGCP agreement has realized separating of control and load plane, and all calling control ' intelligence ' is concentrated on MGC outside the media gateway, and media gateway then is responsible for finishing the processing of audio medium stream.By the MGCP agreement, whole VOIP system realizes that control separates with carrying, has greater flexibility and stronger system processing power.Fig. 3 has shown call control procedure of the present invention:

When (1) user initiated a calling, GW initiated Setup message to MGC, and expression is wished to set up calling, carries caller and called number in Setup message.

(2) after MGC received Setup message, can judgement call accepted, when calling out when legal, just sent the CRCX order to IVR, and notice IVR has the arrival of calling.

(3) after IVR received the CRCX order, the resource that distribution is relevant was replied ACK message then, and time calling is accepted in expression.

(4) after MGC receives ACK, send Alerting message, begin to transmit ring-back tone to GW.

(5) IVR finishes relevant initialization operation, sends NTFY message to MGC, carries the offhook parameter, and expression IVR connects phone.

(6) after MGC receives NTFY message, send Connect message to GW, send MDCX message to IVR simultaneously, call through, caller can stop to the prompt tone of IVR broadcast, and IVR also can begin recording.

(7) IVR starts working according to predefined flow process, and the user can import DTMF button control flow by phone, and the user can import his card number and cipher inquiry remaining sum by phone during for example telephone bank used.GW gives MGC user's DTMF button by UserInput message

(8) MGC gives IVR equipment user key-press message by RQNT message, and IVR enters into corresponding flow process according to user key-press.

(9) user finishes corresponding operating, and after the on-hook, GW sends Release message to MGC.

(10) MGC receives that Release message sends DLCX message, IVR process ends to IVR.

IVR equipment need provide unidirectional voice channel, so that broadcast voice, this just requires IVR equipment can support real-time voice on the IP network, real-time voice on the IP network transmits the general RTP/RTCP of employing agreement at present, IVR equipment then must be supported the RTP/RTCP agreement if need to support plain old telephone directly to visit by visit of IP phone gateway or pc client.At present the voice on the IP network generally all compress with conserve bandwidth, and G.723.1 compression algorithm commonly used have, G.729, can select for use as the case may be.

The present invention carries out the conversion of speech encoding by adopting ACM (Audio Compress Management), solve support to the IP network voice, when receiving, give ACM decoder the relevant information of RTP bag and RTCP bag, the file that is decoded as specified format is preserved; When sending, from the file of preserving in advance, read information, give ACM encoder, restore RTP bag and RTCP and wrap and give the other user.

By the use of ACM, owing to carried out the compression of voice, can significantly reduce the space that file is preserved, by to RTP and RTCP information processing, can provide extraordinary tonequality simultaneously, can satisfy the requirement of voice application fully.

In order to make the operation flow control procedure flexible more and convenient, the present invention adopts the VXML standard.Set up voice VXML document and VXML document management module in interactive responses (IVR) system, the VXML standard is followed in the flow process control of entire I VR system, and by the customization of VXML document.The VXML document of VXML document management module reading system appointment, and explain and carry out the flow process of formulating in the document and control IVR reciprocal process with the user.The advantage of the method is that control is flexible, modification is convenient.VXML mainly contains following syntactic element:

The element title	Chinese is semantic
		??Assign	The value of named variable
??Block	Noninteractive executable code
		??Choice	The definition menu option
??Dtmf	Specify a DFMF
		??Field	Input domain of statement in a form
??Filled	When an input has obtained the action that will carry out after the value
		??Noinput	Catch non-incoming event
??Nomatch	Catch non-match event
		??Prompt	Carry out TTS and voice output to the user
??Record	The record phonetic entry
		??Audio	Play voice
??Throw	Produce an incident

In the control of IVR flow process, mainly contain following requirement:

Recording, playback, reception DTMF input, Text To Speech conversion (Text To Speech:TTS), speech recognition (Auto speech Recognization:ASR), menu are selected and abnormality processing.

Present embodiment requires flow process control among the IVR and the VXML element is set up mapping relations in the following table:

IVR flow process control requirement	The VXML element
		Recording	record
Playback	Audio
		Receive the DTMF input	Dtmf
TTS	prompt
		ASR	Field
Menu is selected	Choice
		Abnormality processing	Throw

By mapping, make all IVR flow process control can represent that all the flow process of writing by the VXML voice of standard just can be supported by IVR equipment like this, has solved the problem of the IVR device intercommunicating of different manufacturers by the VXML voice.

Fig. 5 has further shown the logical construction of IVR of the present invention system, and entire I P IVR system mainly is made up of following components:

IVR Application Logic is the applied logic control section of entire I VR system, and various IVR business externally are provided.

MGCP and RTP/RTCP Stack, this part mainly is responsible for the signaling control and the real-time Transmission of Media Stream on IP network of audio call in system.

VoiceXML Core is core VoiceXML parts, mainly finishes the explanation and the execution of VoiceXML document, is responsible for the control of IVR system business process.Wherein comprise VoiceXML Interpreter and VirtualVoice Device two parts.VoiceXML Interpreter explains and carries out the VoiceXML document.VoiceXML Parser and FIA (Form Interpreter Algorithm) are two piths wherein.Virtual Voice Device is virtual speech ciphering equipment, and it realizes the conversion of VoiceXML speech ciphering equipment to the physics speech ciphering equipment.

IVR Procedure Definition, this is the IVR operation flow and uses relevant definition of object.It comprises two parts, VoiceXML Documents and External VoiceXML Objects, and VoiceXMLDocuments is used to define flow process, and External VoiceXML Objects is and the definition of using relevant VoiceXML external object.Can customize out different IVR business according to user's different demands by this part, make system have higher flexibility.

See also Fig. 6, in IP network, realize the present invention.IP network is based on soft switch (SoftSwitch) framework.Ordinary telephone subscriber is finished visit to the IVR system by IP phone gateway GW by soft switch (SoftSwitch) equipment, the pc user finishes visit to the IVR system by pc client on IP network, Web user then can directly visit the IVR system by browser.

The present invention as call control signalling, has solved the problem with the miscellaneous equipment intercommunication by MGCP, simultaneously because MGCP is comparatively simple, based on MGCP protocol of I VR also can be more or less freely realization.

Claims (4)

1, a kind of method that realizes interactive voice response in the IP network is characterized in that:

Interactive voice response (IVR) system is inserted in the IP network, and described IP network adopts the MGCP agreement that the audio call of interactive voice response system is controlled;

In interactive responses (IVR) system, vocoder is set, the voice that receive are decoded and the voice that send are encoded;

In interactive responses (IVR) system, set up voice VXML document and VXML document management module, and making the operation flow of described VXML document customization and the control flow of interactive responses set up mapping relations, described VXML document management module is responsible for reading, the VXML document of explanation and executive system appointment;

When the user passes through the audio call interactive response system, Media Gateway Controller in the IP network is finished calling signaling reciprocal process and is made the calling party and the interactive response system connection, interactive response system is according to user's demand for services, by the XML document administration module VXML document of system's appointment is operated, with control IVR and user's reciprocal process.

2, method according to claim 1 is characterized in that: described interactive response system inserts IP network by Ethernet interface.

3, method according to claim 1 is characterized in that: described vocoder is ACM decoder and ACM encoder.

4, method according to claim 1 is characterized in that: described mapping relations be the flow process control of interactive responses require and the main syntactic element of VXML standard between mapping relations.

Images