Intelligent voice navigation method, device and system, and computer-readable storage medium

Abstract

The present disclosure relates to an intelligent voice navigation method, device and system, and a computer-readable storage medium. The intelligent voice navigation method comprises: the queuing machine sends the user voice stream to the voice recognition device and the semantic recognition device; the semantic recognition device sends the voice semantic recognition content of the user voice stream to the intelligent voice navigation device; the intelligent voice navigation device generates a short coding identifier according to the voice semantic recognition content, and establishes a corresponding relationship between the voice semantic recognition content and the short coding identifier; the intelligent voice navigation device returns the short coding identifier to the queuing machine; the queuing machine sends the short coding identifier to the interactive voice response server; the interactive voice response server obtains the corresponding voice semantic recognition content from the intelligent voice navigation device according to the short coding identifier. The present disclosure is based on a custom XID short code mechanism, which solves the problem that mainstream queuing machines are limited in recognizing too long text information for intelligent navigation.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence, and in particular to an intelligent voice navigation method, device and system, and a computer-readable storage medium. Background Art

[0002] Related Technology Many call center networks currently use the IVR (Interactive Voice Response) process implementation method, which is to transmit the user's key information (code) to the IVR server through the voice queuing machine, and the IVR server drives the next action. Therefore, the data transmitted between the queuing machine and the IVR server does not need to be defined to be very long. For example, Huawei's call center only supports 23 bytes, that is, a maximum of 11 Chinese characters.

[0003] To implement voice intelligent navigation in such a situation, it is necessary to transmit interactive semantic information obtained from the analysis of user voice. However, this identification information often exceeds the byte limit of the queue machine and cannot be transmitted to the IVR server for process node execution.

[0004] This situation exists in call centers built before 2016, including Huawei UAP8100, UAP3100, ZTE MS10 and other models, which makes it difficult for traditional call centers to implement intelligent voice navigation. The solutions of related technologies require major adjustments to the IVR process code mechanism. Summary of the invention

[0005] In view of at least one of the above technical problems, the present disclosure provides an intelligent voice navigation method, device and system, and a computer-readable storage medium, which, based on a custom XID short code mechanism, solves the problem that mainstream queuing machines are limited in recognizing too long text information for intelligent navigation.

[0006] According to one aspect of the present disclosure, there is provided an intelligent voice navigation method, comprising:

[0007] The queuing machine sends the user voice stream to the voice recognition device and the semantic recognition device;

[0008] The semantic recognition device sends the speech semantic recognition content of the user's voice stream to the intelligent voice navigation device;

[0009] The intelligent voice navigation device generates a short code identifier according to the voice semantic recognition content, and establishes a corresponding relationship between the voice semantic recognition content and the short code identifier;

[0010] The intelligent voice navigation device returns the short code identification to the queuing machine;

[0011] The queuing machine sends the short code identification to the interactive voice response server;

[0012] The interactive voice response server obtains corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier.

[0013] In some embodiments of the present disclosure, the interactive voice response server obtains corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier, including:

[0014] The intelligent voice navigation device obtains the corresponding voice semantic recognition content according to the short code identifier;

[0015] The intelligent voice navigation device generates a corresponding interactive voice process execution node according to the voice semantic recognition content;

[0016] The intelligent voice navigation device encapsulates the interactive voice process execution node into a process node voice extensible markup language that can be executed by an interactive voice response server;

[0017] The intelligent voice navigation device returns the process node voice extensible markup language to the interactive voice response server.

[0018] In some embodiments of the present disclosure, the intelligent voice navigation method further includes:

[0019] For scenarios with multiple rounds of interactions in intelligent voice navigation, the intelligent voice navigation device cooperates with the interactive voice response server to convert the voice semantic recognition content into short-coded identifiers, and establishes the association logic of the semantic context to achieve multi-semantic slot completion recognition.

[0020] In some embodiments of the present disclosure, the queuing machine sending the user voice stream to the voice recognition device and the semantic recognition device includes:

[0021] The queuing machine sends the user's voice stream to the interactive voice response server to request voice distribution;

[0022] The interactive voice response server instructs the queuing machine to send the user voice stream to the voice recognition device;

[0023] The speech recognition device performs speech recognition on the user's speech stream and sends the text after speech recognition to the semantic recognition device.

[0024] According to another aspect of the present disclosure, there is provided an intelligent voice navigation device, comprising:

[0025] The short code conversion module is used to receive the speech and semantic recognition content of the user's voice stream, wherein the queuing machine sends the user's voice stream to the voice recognition device and the semantic recognition device, and the semantic recognition device outputs the speech and semantic recognition content of the user's voice stream to the intelligent voice navigation device; generates a short code identifier according to the speech and semantic recognition content, establishes a corresponding relationship between the speech and semantic recognition content and the short code identifier; and returns the short code identifier to the queuing machine, instructing the queuing machine to send the short code identifier to the interactive voice response server;

[0026] The short code decoding module is used to parse the short code identifier of the interactive voice response server, obtain the speech semantic recognition content corresponding to the short code identifier; and return the speech semantic recognition content corresponding to the short code identifier to the interactive voice response server.

[0027] In some embodiments of the present disclosure, a short code decoding module is used to obtain corresponding speech semantic recognition content according to a short code identifier; generate a corresponding interactive speech process execution node according to the speech semantic recognition content; encapsulate the interactive speech process execution node into a process node speech extensible markup language that can be executed by an interactive voice response server; and return the process node speech extensible markup language to the interactive voice response server.

[0028] In some embodiments of the present disclosure, a short code conversion module and a short code decoding module are used to cooperate with an interactive voice response server for scenarios of multiple rounds of interaction in intelligent voice navigation, and to establish association logic of the semantic context when performing short code identification conversion on the voice semantic recognition content, so as to realize multi-semantic slot completion recognition.

[0029] In some embodiments of the present disclosure, the intelligent voice navigation device further includes:

[0030] The message middleware module is used to store the correspondence between speech semantic recognition content and short code identifier.

[0031] According to another aspect of the present disclosure, there is provided an intelligent voice navigation system, comprising:

[0032] A queuing machine, used for sending the user voice stream to the voice recognition device and the semantic recognition device;

[0033] A semantic recognition device, used for sending the speech semantic recognition content of the user's voice stream to the intelligent voice navigation device;

[0034] The intelligent voice navigation device is used to generate a short code identifier according to the voice semantic recognition content, establish a corresponding relationship between the voice semantic recognition content and the short code identifier; return the short code identifier to the queuing machine, and instruct the queuing machine to send the short code identifier to the interactive voice response server;

[0035] The interactive voice response server is used to obtain corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier.

[0036] In some embodiments of the present disclosure, the intelligent voice navigation system further includes:

[0037] The queuing machine is also used to send the user's voice stream to the interactive voice response server to request voice distribution;

[0038] The interactive voice response server is also used to instruct the queuing machine to send the user voice stream to the voice recognition device;

[0039] The speech recognition device is used to perform speech recognition on the user's speech stream and send the text after speech recognition to the semantic recognition device.

[0040] In some embodiments of the present disclosure, the intelligent voice navigation device is an intelligent voice navigation device as described in any of the above embodiments.

[0041] According to another aspect of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and when the instructions are executed by a processor, the intelligent voice navigation method as described in any of the above embodiments is implemented.

[0042] The present invention is based on a custom XID short code mechanism, which solves the problem that mainstream queuing machines are limited in their ability to recognize text messages that are too long for intelligent navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 Schematic diagram of some embodiments of the intelligent voice navigation method disclosed herein.

[0045] Figure 2 Schematic diagrams of other embodiments of the intelligent voice navigation method disclosed herein.

[0046] Figure 3 It is a schematic diagram of a short code identification multi-round interaction mechanism in some embodiments of the present disclosure.

[0047] Figure 4 Schematic diagram of some embodiments of the intelligent voice navigation device disclosed herein.

[0048] Figure 5Schematic diagram of some embodiments of the intelligent voice navigation system disclosed herein. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present disclosure and its application or use. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0050] Unless specifically stated otherwise, the relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure.

[0051] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0052] Technologies, methods, and apparatus known to ordinary technicians in the relevant field may not be discussed in detail, but where appropriate, such technologies, methods, and apparatus should be considered part of the authorization specification.

[0053] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0054] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0055] Figure 1 Schematic diagram of some embodiments of the intelligent voice navigation method of the present disclosure. Preferably, this embodiment can be performed by the intelligent voice navigation device of the present disclosure or the intelligent voice navigation system of the present disclosure. The method may include steps 11 to 16, wherein:

[0056] Step 11: The queuing machine sends the user voice stream to the voice recognition device and the semantic recognition device.

[0057] In some embodiments of the present disclosure, step 11 may include steps 111 to 113, wherein:

[0058] Step 111: The queuing machine sends the user voice stream to the interactive voice response server to request voice distribution.

[0059] Step 112: The interactive voice response server instructs the queuing machine to send the user voice stream to the voice recognition device.

[0060] Step 113: the speech recognition device performs speech recognition on the user's speech stream and sends the text after speech recognition to the semantic recognition device.

[0061] Step 12: The semantic recognition device sends the speech semantic recognition content of the user's voice stream to the intelligent voice navigation device.

[0062] Step 13, the intelligent voice navigation device generates a short code identifier XID according to the voice semantic recognition content, and uses a Redis (Remote Dictionary Server) memory server to establish a corresponding relationship between the voice semantic recognition content and the short code identifier.

[0063] In some embodiments of the present disclosure, Redis is a mature, open source data cache service architecture that is widely used for fast memory-based data reading and writing. Redis clusters can be used to achieve efficient real-time message stack management. The above embodiments of the present disclosure use Redis to build a transmission schedule that carries highly concurrent intelligent navigation interactive semantic XML data messages.

[0064] Step 14: The intelligent voice navigation device returns the short code identifier to the queuing machine.

[0065] Step 15: The queuing machine sends the short code identifier to the interactive voice response server.

[0066] Step 16: The interactive voice response server obtains corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier.

[0067] In some embodiments of the present disclosure, step 16 may include steps 161 to 164, wherein:

[0068] Step 161: The intelligent voice navigation device obtains corresponding voice semantic recognition content according to the short code identifier.

[0069] Step 162: The intelligent voice navigation device generates a corresponding interactive voice process execution node according to the voice semantic recognition content.

[0070] Step 163: the intelligent voice navigation device encapsulates the interactive voice process execution node into a process node voice extensible markup language that can be executed by an interactive voice response server.

[0071] Step 164: the intelligent voice navigation device returns the process node voice extensible markup language to the interactive voice response server.

[0072] In some embodiments of the present disclosure, the intelligent voice navigation method may further include: for scenarios with multiple rounds of interaction in intelligent voice navigation, the intelligent voice navigation device cooperates with an interactive voice response server to perform short code identification conversion on the voice semantic recognition content, and establishes association logic for the semantic context to achieve multi-semantic slot completion recognition.

[0073] Based on the intelligent voice navigation method provided by the above embodiment of the present disclosure, the original Code mechanism and core architecture are not changed. The above embodiment of the present disclosure uses the Redis memory server, customizes the XID short code, and establishes a temporary relationship with the user's voice recognition text, thereby solving the problem of limited data transmission length for intelligent navigation voice semantic analysis and recognition, avoiding the large amount of investment brought by the call center upgrade, and accelerating the launch of intelligent voice navigation.

[0074] The above-mentioned embodiments of the present disclosure can implement intelligent IVR navigation services based on traditional call centers, and support multiple rounds of interaction between customer service robots and users, thereby replacing traditional key-based IVR services and achieving flattening.

[0075] Figure 2 Schematic diagrams of other embodiments of the intelligent voice navigation method of the present disclosure. Preferably, this embodiment can be performed by the intelligent voice navigation device of the present disclosure or the intelligent voice navigation system of the present disclosure. The method may include steps 201 to 216, wherein:

[0076] Step 201: When a user makes a phone call to a queuing machine through a user terminal, the queuing machine sends the user's voice stream to an interactive voice response (IVR) server to request voice distribution.

[0077] Step 202, the interactive voice response server returns the guiding words to the queuing machine, instructing the queuing machine to send the user voice stream to the ASR (Automatic Speech Recognition, also known as speech recognition device).

[0078] Step 203: the queuing machine sends the user voice stream to the voice recognition device.

[0079] Step 204, the speech recognition device performs speech recognition on the user's speech stream, and sends the text after speech recognition to the semantic NLP (Natural Language Processing, also known as semantic recognition device).

[0080] Step 205: the semantic recognition device sends the speech semantic recognition content of the user's voice stream to the intelligent voice navigation device.

[0081] In some embodiments of the present disclosure, the speech semantic recognition content of the user voice stream may be a semantic recognition result of a semantic recognition device, and then step 206 and step 207 are performed.

[0082] In some embodiments of the present disclosure, the speech semantic recognition content of the user voice stream may be interactive semantic XML (eXtensible Markup Language).

[0083] Step 206 , the short code conversion module of the intelligent voice navigation device generates a short code identifier XID according to the voice semantic recognition content, and returns the short code identifier to the queuing machine, and then executes step 208 .

[0084] Step 207: the intelligent voice navigation device uses the Redis service to cache the speech semantic recognition content of the user's voice stream, and establishes a corresponding relationship between the speech semantic recognition content and the short code identifier.

[0085] Step 208: The queuing machine transparently transmits the short code identifier XID to the interactive voice response server.

[0086] Step 209: the interactive voice response server requests the node to execute and transparently transmit the short code identifier XID to the short code decoding module of the intelligent voice navigation device.

[0087] Step 210, the short code decoding module of the intelligent voice navigation device parses the short code identifier XID and requests the speech semantic recognition content corresponding to the short code identifier, that is, requests the speech semantic recognition content (such as interactive semantic XML) corresponding to the short code identifier from the Redis message middleware module.

[0088] Step 211, the Redis message middleware module of the intelligent voice navigation device returns the speech semantic recognition content (such as interactive semantic XML) corresponding to the short code identifier to the short code decoding module.

[0089] Step 212, the short code decoding module of the intelligent voice navigation device generates a corresponding interactive voice process execution node according to the voice semantic recognition content; encapsulates the interactive voice process execution node into a process node VXML (Voice eXtensible Markup Language) that can be executed by an interactive voice response server; and returns the process node VXML to the interactive voice response server.

[0090] In some embodiments of the present disclosure, step 212 may include: returning a complete node VXML in response to the IVR process request.

[0091] In some embodiments of the present disclosure, the IVR server uses steps 209-212 to obtain real user voice and semantic recognition content in the intelligent voice navigation device through XID, thereby driving the subsequent actions of the queuing machine (steps 213-216).

[0092] Step 213: the interactive voice response server initiates an IVR process node and returns the voice distribution result to the queuing machine.

[0093] Step 214: the queuing machine requests a TTS (Text To Speech) speech synthesis device to broadcast a TTS announcement.

[0094] Step 215: The TTS speech synthesis device synthesizes the announcement sound and returns it to the queuing machine.

[0095] Step 216: the queuing machine broadcasts the announcement tone to the user.

[0096] The above-mentioned embodiment of the present disclosure provides an intelligent voice navigation integration method based on dynamically generated XID short code data by establishing an adapter device with a Redis cache service, so as to solve the problem that the originally complex recognition text information whose byte number exceeds the limit of the queue machine cannot be transmitted, and finally realizes the clear mapping of the user's voice interaction recognition information to the voice navigation IVR process execution node.

[0097] The voice stream output by the queuing machine in the above embodiment of the present disclosure is converted through an interface protocol and sent to a voice recognition and semantic recognition engine. The recognized content is sent to this device for caching through the Redis service, and an XID short code is dynamically generated to establish an association relationship.

[0098] The intelligent voice navigation device of the above embodiment of the present disclosure sends the XID to the queuing machine, and the queuing machine sends the XID to the IVR server, thereby bypassing the limitations of the traditional queuing machine that does not support voice recognition content and has limited transmission data.

[0099] The above-mentioned embodiment of the present disclosure solves the problem that mainstream queuing machines are limited in recognizing too long text information for intelligent navigation based on a custom XID short code mechanism.

[0100] The above-mentioned embodiments of the present disclosure can achieve the de-MRCP (Media Resource Control Protocol) protocol architecture by packet capture, breaking through the limitation that traditional voice AI engines need to be privately deployed. Based on the new architecture, the above-mentioned embodiments of the present disclosure can adapt to the mainstream voice AI engine interface protocol. The engine only needs to provide a standardized and open http (HyperText Transfer Protocol) or hypertext transfer protocol / websocket (a protocol for full-duplex communication on a single TCP connection) interface.

[0101] Figure 3 The figure is a schematic diagram of the short code identification multi-round interaction mechanism in some embodiments of the present disclosure. For multi-round interactions in intelligent voice navigation, the intelligent voice navigation device of the present disclosure cooperates with the IVR server to provide a multi-semantic slot completion recognition mechanism, which can carry the context association logic when converting the voice recognition data into XID code, so as to better support the scenario of multi-round human-computer dialogue in one business.

[0102] For example: Figure 3 In the embodiment, the bill checking code is determined according to the voice stream of the user "I want to check the bill", and the intermediate node is determined to be 01002, which is used as the starting slot in the multi-semantic slots. The queuing machine broadcasts the voice of "Confirm whether the query is for this machine or another machine" to the user, and checks the code of this machine according to the voice stream of "my machine" returned by the user, and the intermediate node + 01 is used as the middle slot in the multi-semantic slots. The queuing machine broadcasts the voice of "Excuse me, which month of this year do you need to check the bill for?" to the user, and checks the code of February according to the voice stream of "February" returned by the user, and the intermediate node + 02 is used as the ending slot in the multi-semantic slots. The semantics of the three slots of February, the machine, and the bill are integrated for contextual association, and finally the total bill amount of the user in February is provided to the user, and the user is reminded to pay.

[0103] The above-mentioned embodiments of the present disclosure can implement a multi-semantic slot completion and recognition mechanism for XID short codes based on session association, thereby realizing multiple rounds of human-computer dialogue for one business.

[0104] The above-mentioned embodiments of the present invention disclose AI (Artificial Intelligence) processing and analysis of real-time traffic on the call platform. Compared with the traditional customized development of traffic logic at the application level in related technologies, it can realize dual-channel traffic information management and real-time integration in the intelligent voice navigation device, and can realize the associated output of traffic route and voice stream AI processing results, with high reusability value.

[0105] Figure 4 Schematic diagram of some embodiments of the intelligent voice navigation device disclosed in the present invention. Figure 4 As shown, the intelligent voice navigation device of the present disclosure may include a short code conversion module 41 and a short code decoding module 42, wherein:

[0106] The short code conversion module 41 is used to receive the speech and semantic recognition content of the user voice stream, wherein the queuing machine sends the user voice stream to the speech recognition device and the semantic recognition device, and the semantic recognition device outputs the speech and semantic recognition content of the user voice stream to the intelligent voice navigation device; generates a short code identifier according to the speech and semantic recognition content, establishes a corresponding relationship between the speech and semantic recognition content and the short code identifier; and returns the short code identifier to the queuing machine, instructing the queuing machine to send the short code identifier to the interactive voice response server.

[0107] In some embodiments of the present disclosure, the short code conversion module 41 may be a semantic XID short code conversion module.

[0108] In some embodiments of the present disclosure, the short code conversion module 41 can be used to define the XID short code based on the node execution rule, and replace the original semantic recognition output result with too many words with the corresponding code, thereby bypassing the queue machine limitation.

[0109] The short code decoding module 42 is used to parse the short code identifier of the interactive voice response server, obtain the speech semantic recognition content corresponding to the short code identifier, and return the speech semantic recognition content corresponding to the short code identifier to the interactive voice response server.

[0110] In some embodiments of the present disclosure, the short code decoding module 42 can be used to obtain corresponding speech semantic recognition content based on the short code identifier; generate a corresponding interactive speech process execution node based on the speech semantic recognition content; encapsulate the interactive speech process execution node into a process node speech extensible markup language that can be executed by an interactive voice response server; and return the process node speech extensible markup language to the interactive voice response server.

[0111] In some embodiments of the present disclosure, the short code conversion module 41 may be an XID short code node parsing module.

[0112] In some embodiments of the present disclosure, the short code conversion module 41 can be used to parse the generated XID short code, obtain multiple corresponding IVR process execution nodes, and encapsulate them into a process node VXML that can be executed by the IVR.

[0113] In some embodiments of the present disclosure, the short code conversion module 41 and the short code decoding module 42 can be used for scenarios of multiple rounds of interaction in intelligent voice navigation, and cooperate with an interactive voice response server to perform short code identification conversion on the voice semantic recognition content, and establish association logic of the semantic context to achieve multi-semantic slot completion recognition.

[0114] In some embodiments of the present disclosure, Figure 4 As shown, the intelligent voice navigation device may further include a message middleware module 43, wherein:

[0115] The message middleware module 43 is used to store the corresponding relationship between the speech semantic recognition content and the short code identifier.

[0116] In some embodiments of the present disclosure, the message middleware module 43 may be a Redis message middleware module.

[0117] In some embodiments of the present disclosure, the message middleware module 43 may be used to carry the transmission scheduling of high-concurrency intelligent navigation interaction semantic XML data messages.

[0118] Based on the intelligent voice navigation device provided by the above embodiment of the present disclosure, the original Code mechanism and core architecture are not changed. The above embodiment of the present disclosure uses the Redis memory server, customizes the XID short code, and establishes a temporary relationship with the user's voice recognition text, thereby solving the problem of limited data transmission length for intelligent navigation voice semantic analysis and recognition, avoiding the large investment caused by call center upgrades, and accelerating the launch of intelligent voice navigation.

[0119] The above-mentioned embodiments of the present disclosure can implement intelligent IVR navigation services based on traditional call centers, and support multiple rounds of interaction between customer service robots and users, thereby replacing traditional key-based IVR services and achieving flattening.

[0120] Figure 5 Schematic diagram of some embodiments of the intelligent voice navigation system disclosed in the present invention. Figure 5 As shown, the intelligent voice navigation device disclosed in the present invention may include a queuing machine 1, a semantic recognition device 2, an interactive voice response server 3 and an intelligent voice navigation device 4, wherein:

[0121] The queuing machine 1 is used to send the user voice stream to the voice recognition device and the semantic recognition device.

[0122] The semantic recognition device 2 is used to send the speech semantic recognition content of the user's voice stream to the intelligent voice navigation device.

[0123] In some embodiments of the present disclosure, the speech semantic recognition content of the user voice stream may be interactive semantic XML.

[0124] The intelligent voice navigation device 4 is used to generate a short code identifier XID according to the voice semantic recognition content, establish a corresponding relationship between the voice semantic recognition content and the short code identifier; return the short code identifier to the queuing machine, and instruct the queuing machine to send the short code identifier to the interactive voice response server.

[0125] The interactive voice response server 3 is used to obtain the corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier.

[0126] In some embodiments of the present disclosure, Figure 5 As shown, the intelligent voice navigation system may further include a voice recognition device 5, wherein:

[0127] The queuing machine 1 can also be used to send the user's voice stream to the interactive voice response IVR server to request voice distribution when a user calls into the queuing machine through a user terminal.

[0128] The interactive voice response server 3 can also be used to return the guiding words to the queuing machine, instructing the queuing machine to send the user voice stream to the voice recognition device.

[0129] The speech recognition device 5 is used to perform speech recognition on the user's speech stream and send the text after speech recognition to the semantic recognition device 2.

[0130] The semantic recognition device 2 is used to send the speech semantic recognition content of the user voice stream to the intelligent voice navigation device 4.

[0131] In some embodiments of the present disclosure, Figure 5 As shown, the speech semantic recognition content may be a semantic result encapsulated in XML format.

[0132] In some embodiments of the present disclosure, Figure 5 As shown, the intelligent voice navigation device 4 may include a short code conversion module 41, a short code decoding module 42 and a message middleware module 43, wherein:

[0133] The short code conversion module 41 of the intelligent voice navigation device 4 is used to generate a short code identifier XID according to the voice semantic recognition content (semantic XML), establish a corresponding relationship between the voice semantic recognition content and the short code identifier in the short code decoding module 42; return the short code identifier to the queuing machine 1, and instruct the queuing machine 1 to send the short code identifier to the interactive voice response server.

[0134] The interactive voice response server 3 is used for requesting the node to execute and transparently transmit the short code identifier XID to the short code decoding module 42 of the intelligent voice navigation device.

[0135] The short code decoding module 42 of the intelligent voice navigation device is used to parse the short code identifier XID and request the speech semantic recognition content corresponding to the short code identifier, that is, to request the speech semantic recognition content (such as interactive semantic XML) corresponding to the short code identifier from the Redis message middleware module.

[0136] The Redis message middleware module 43 of the intelligent voice navigation device is used to return the voice semantic recognition content (such as interactive semantic XML) corresponding to the short code identifier to the short code decoding module 42.

[0137] The short encoding decoding module 42 of the intelligent voice navigation device is used to generate a corresponding interactive voice process execution node (such as an XML encapsulation node) according to the voice semantic recognition content; encapsulate the interactive voice process execution node into a process node VXML that can be executed by an interactive voice response server; and return the process node voice extensible markup language to the interactive voice response server.

[0138] Figure 2 Schematic diagrams of other embodiments of the intelligent voice navigation system disclosed in the present invention are also provided. Figure 5 Compared with the embodiment, Figure 2 The intelligent voice navigation system of the embodiment includes not only a queuing machine 1, a voice recognition device 5, a semantic recognition device 2, an interactive voice response server 3 and an intelligent voice navigation device 4, but also a TTS voice synthesis device 6, wherein:

[0139] The interactive voice response server 3 can also be used to initiate an IVR process node and return the voice distribution result to the queuing machine.

[0140] The queuing machine 1 can also be used to request a TTS announcement from the TTS speech synthesis device.

[0141] The TTS speech synthesis device 6 is used to synthesize the announcement sound and return it to the queuing machine 1, and instruct the queuing machine 1 to broadcast the announcement sound to the user.

[0142] In some embodiments of the present disclosure, Figure 2 and Figure 3 The intelligent voice navigation device of the embodiment may be any of the above embodiments (eg Figure 4 The intelligent voice navigation device described in embodiment).

[0143] Based on the intelligent voice navigation system provided by the above embodiment of the present disclosure, the original Code mechanism and core architecture are not changed. The above embodiment of the present disclosure uses the Redis memory server, customizes the XID short code, and establishes a temporary relationship with the user's voice recognition text, thereby solving the problem of limited data transmission length for intelligent navigation voice semantic analysis and recognition, avoiding the large investment brought by the call center upgrade, and accelerating the launch of intelligent voice navigation.

[0144] The above-mentioned embodiments of the present disclosure can implement intelligent IVR navigation services based on traditional call centers, and support multiple rounds of interaction between customer service robots and users, thereby replacing traditional key-based IVR services and achieving flattening.

[0145] The intelligent voice navigation method, device and system adopted in the above embodiments of the present disclosure are actually used in a certain operator's multimedia customer service cooperation group intelligent IVR flattening construction project. The above embodiments of the present disclosure are verified in this project, which realizes the rapid launch of the intelligent voice navigation service of a certain city operator and avoids the transformation and upgrading of the call center.

[0146] According to another aspect of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and when the instructions are executed by a processor, any of the above embodiments (for example Figure 1 or Figure 2 The intelligent voice navigation method described in embodiment).

[0147] Based on the computer-readable storage medium provided by the above embodiment of the present disclosure, the original Code mechanism and core architecture are not changed. The above embodiment of the present disclosure uses the Redis memory server, customizes the XID short code, and establishes a temporary relationship with the user's voice recognition text, thereby solving the problem of limited data transmission length for intelligent navigation voice semantic analysis and recognition, avoiding the large amount of investment brought by the call center upgrade, and accelerating the launch of intelligent voice navigation.

[0148] The above-mentioned embodiments of the present disclosure can implement intelligent IVR navigation services based on traditional call centers, and support multiple rounds of interaction between customer service robots and users, thereby replacing traditional key-based IVR services and achieving flattening.

[0149] The voice stream output by the queuing machine in the above embodiment of the present disclosure is converted through an interface protocol and sent to a voice recognition and semantic recognition engine. The recognized content is sent to this device for caching through the Redis service, and an XID short code is dynamically generated to establish an association relationship.

[0150] The intelligent voice navigation device of the above embodiment of the present disclosure sends the XID to the queuing machine, and the queuing machine sends the XID to the IVR server, thereby bypassing the limitations of the traditional queuing machine that does not support voice recognition content and has limited transmission data.

[0151] The above-mentioned embodiment of the present disclosure solves the problem that mainstream queuing machines are limited in recognizing too long text information for intelligent navigation based on a custom XID short code mechanism.

[0152] The above-mentioned embodiments of the present disclosure can realize the de-MRCP protocol architecture by packet capture, breaking through the limitation that the traditional voice AI engine needs to be privately deployed. Based on the new architecture, the above-mentioned embodiments of the present disclosure can adapt to the mainstream voice AI engine interface protocol. The engine only needs to provide a standardized and open http (HyperText Transfer Protocol) or hypertext transfer protocol / websocket (a protocol for full-duplex communication on a single TCP connection) interface.

[0153] The intelligent voice navigation device described above can be implemented as a general-purpose processor, a programmable logic controller (PLC), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components or any appropriate combination thereof for performing the functions described in this application.

[0154] So far, the present disclosure has been described in detail. In order to avoid obscuring the concept of the present disclosure, some details known in the art are not described. Based on the above description, those skilled in the art can fully understand how to implement the technical solution disclosed here.

[0155] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or may be accomplished by instructing the relevant hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0156] The description of the present disclosure is given for the purpose of illustration and description, and is not intended to be exhaustive or to limit the present disclosure to the disclosed form. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present disclosure, and to enable those of ordinary skill in the art to understand the present disclosure and thereby design various embodiments with various modifications suitable for specific uses.

Claims

1. An intelligent voice navigation method, characterized in that: include: The queuing machine sends the user voice stream to the voice recognition device and the semantic recognition device; The semantic recognition device sends the speech semantic recognition content of the user's voice stream to the intelligent voice navigation device; The intelligent voice navigation device generates a short code identifier according to the voice semantic recognition content, and establishes a corresponding relationship between the voice semantic recognition content and the short code identifier; The intelligent voice navigation device returns the short code identification to the queuing machine; The queuing machine sends the short code identification to the interactive voice response server; The interactive voice response server obtains the corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier; The interactive voice response server obtains the corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier, including: The intelligent voice navigation device obtains the corresponding voice semantic recognition content according to the short code identifier; The intelligent voice navigation device generates a corresponding interactive voice process execution node according to the voice semantic recognition content; The intelligent voice navigation device encapsulates the interactive voice process execution node into a process node voice extensible markup language that can be executed by an interactive voice response server; The intelligent voice navigation device returns the process node voice extensible markup language to the interactive voice response server.

2. The intelligent voice navigation method according to claim 1, characterized in that: Also includes: For scenarios with multiple rounds of interactions in intelligent voice navigation, the intelligent voice navigation device cooperates with the interactive voice response server to convert the voice semantic recognition content into short-coded identifiers, and establishes the association logic of the semantic context to achieve multi-semantic slot completion recognition.

3. The intelligent voice navigation method according to claim 1 or 2, characterized in that: The queuing machine sends the user voice stream to the voice recognition device and the semantic recognition device, including: The queuing machine sends the user's voice stream to the interactive voice response server to request voice distribution; The interactive voice response server instructs the queuing machine to send the user voice stream to the voice recognition device; The speech recognition device performs speech recognition on the user's speech stream and sends the text after speech recognition to the semantic recognition device.

4. An intelligent voice navigation device, characterized in that: include: The short code conversion module is used to receive the speech and semantic recognition content of the user's voice stream, wherein the queuing machine sends the user's voice stream to the voice recognition device and the semantic recognition device, and the semantic recognition device outputs the speech and semantic recognition content of the user's voice stream to the intelligent voice navigation device; generates a short code identifier according to the speech and semantic recognition content, establishes a corresponding relationship between the speech and semantic recognition content and the short code identifier; and returns the short code identifier to the queuing machine, instructing the queuing machine to send the short code identifier to the interactive voice response server; The short code decoding module is used to parse the short code identifier of the interactive voice response server, obtain the speech semantic recognition content corresponding to the short code identifier, and return the speech semantic recognition content corresponding to the short code identifier to the interactive voice response server; Among them, the short code decoding module is used to obtain the corresponding speech semantic recognition content according to the short code identifier; generate the corresponding interactive speech process execution node according to the speech semantic recognition content; encapsulate the interactive speech process execution node into a process node speech extensible markup language that can be executed by an interactive voice response server; and return the process node speech extensible markup language to the interactive voice response server.

5. The intelligent voice navigation device according to claim 4, characterized in that: The short code conversion module and the short code decoding module are used to cooperate with the interactive voice response server for multi-round interaction scenarios in intelligent voice navigation, and establish the association logic of the semantic context when converting the voice semantic recognition content into short code identification to achieve multi-semantic slot completion recognition.

6. The intelligent voice navigation device according to claim 4 or 5, characterized in that: Also includes: The message middleware module is used to store the correspondence between speech semantic recognition content and short code identifier.

7. An intelligent voice navigation system, characterized in that: include: A queuing machine, used for sending the user voice stream to the voice recognition device and the semantic recognition device; A semantic recognition device, used for sending the speech semantic recognition content of the user's voice stream to the intelligent voice navigation device; An intelligent voice navigation device, used to generate a short code identifier according to the voice semantic recognition content, and establish a corresponding relationship between the voice semantic recognition content and the short code identifier; Returning the short code identifier to the queuing machine, and instructing the queuing machine to send the short code identifier to the interactive voice response server; An interactive voice response server, used to obtain corresponding voice semantic recognition content from the intelligent voice navigation device according to the short code identifier; Among them, the intelligent voice navigation device is used to obtain the corresponding voice semantic recognition content according to the short code identifier; generate the corresponding interactive voice process execution node according to the voice semantic recognition content; encapsulate the interactive voice process execution node into a process node voice extensible markup language that can be executed by an interactive voice response server; and return the process node voice extensible markup language to the interactive voice response server.

8. The intelligent voice navigation system according to claim 7, characterized in that: Also includes: The queuing machine is also used to send the user's voice stream to the interactive voice response server to request voice distribution; The interactive voice response server is also used to instruct the queuing machine to send the user voice stream to the voice recognition device; The speech recognition device is used to perform speech recognition on the user's speech stream and send the text after speech recognition to the semantic recognition device.

9. The intelligent voice navigation system according to claim 7 or 8, characterized in that: The intelligent voice navigation device is used to cooperate with an interactive voice response server in scenarios of multiple rounds of interaction in intelligent voice navigation, and to establish association logic of semantic context when converting voice semantic recognition content into short code identifiers, so as to realize multi-semantic slot completion recognition.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the instructions are executed by a processor, the intelligent voice navigation method according to any one of claims 1 to 3 is implemented.

Images