JPH03159356A - Response method for audio response recognizing device - Google Patents

Abstract

PURPOSE:To efficiently progress service by automatically selecting the necessity of plural response guidance or a sign tone sending function corresponding to a user. CONSTITUTION:The audio response recognizing device is comprised of a network control part 1, a speech recognition part 2, an audio response part 3, and a main control part 4, and when an outgoing talephone number is inputted from the user, the degree of operating skill for a system of the user prepared as a table in the device is discriminated based on the telephone number. Then, the response guidance is selected based on discriminated degree of operating skill, and also, it is selected whether or not a sign tone to expedite the user to be easy to perform voice input based on procedure shown by the response guidance is sent. Thereby, it is possible to efficiently progress audio response service corresponding to a wide range of users.

Description

[Detailed description of the invention]

[Industrial Application Field] The present invention relates to a voice response recognition device used for bank balance inquiry and transfer notification services, etc., which identifies the user's voice and automatically responds while proceeding with the service. In particular, networks with calling phone number notification functionality, e.g.
ISDN (Integrated Serv
The present invention relates to a response method of a voice response recognition device having an interface connectable to the ICES Digital Network (hereinafter referred to as integrated digital communication network). [Prior Art] Voice response recognition devices that use voice recognition technology for non-specific speakers are widely commercialized and are used for bank balance inquiries and transfer notifications, stock price inquiries at securities companies, etc. In conventional systems, interaction with voice response recognition devices uses response guidance that is fixed for each service. In addition, in order to perform voice input at accurate timing, a signal sound that prompts vocalization is fixedly output when the voice input is triggered. FIG. 6 is an explanatory diagram showing service specifications in response to voice input in a conventional system. Here, the case of inputting a service code is shown. In the conventional system, the voice response recognition device uniformly outputs the guidance voice/signal sound 6l as shown in this example.
Message 63 "Please enter your PIN number" is used,
Signal sounds 64 and 66 are output. The user inputs a predetermined input rOJ 65. after each signal sound as the input voice 62. Do NJ 67. A message 68 confirming the user's input is displayed as ``Balance Inquiry. First, the voice response device outputs the guidance 7l (“Please make a service call”), then the signal sound 72 (“beep”) is output, and immediately after that, the voice recognition device starts the voice recognition operation 74. do. The speech recognition operation 74 continues for a certain period of time, for example, about two seconds in the case of mundane, and recognizes the speech input 73 uttered within that period. In this way, the speech recognition device receives a processing request from a user. On the other hand, in ISDN networks and the like, the caller's telephone number can be notified to the other party as a network function. This caller's phone number does not represent the caller itself, but
In the case of a general household, the caller can be thought of as a fixed caller, temporarily and approximately. ISDN connects transmission lines and exchanges with digital lines,
It is a public digital network that provides various communication services such as voice, data, and images through a unified interface. In addition to integrating telephone, facsimile, videotex, and data communications into a single subscriber line, the service provides services that enable image communications such as videophone calls. Regarding the application of ISDN, please refer to ``Guidelines for Utilization of TSDtJ for Nikkei Communication'' published by Nikkei BP, especially the above text is excerpted from PP.18 and PP.189.

[Problem to be solved by the invention]

The voice response recognition device in the conventional system uses fixed response guidance for each service or outputs a fixed signal sound to encourage speaking. However, it is desirable to be able to respond flexibly to the user's response guidance and signal sound transmission function, depending on the user's system proficiency level, the voice recognition performance expected of the user, and so on. In other words, in the conventional system, a signal sound is used to give an opportunity to input the voice to be recognized, but if the message before the signal sound is simply ``R PIN number please J'', it is difficult to get used to it. People often make the mistake of starting to speak before the signal sound is output, or overlapping the sound with the signal sound. Further, if the signal sound is output after the audio guidance indicating that the user wants to speak after the signal sound is output, it becomes redundant for people who are used to the signal sound, which is inconvenient. Furthermore, for those who are accustomed to the system, it is no longer necessary to use the cue sound itself, and it is hoped that they will be able to respond more efficiently. For example, as shown in FIGS. 6 and 7, in the conventional system, beginners may not be able to properly grasp the timing. For such beginners, it is easier to input the password by saying, ``Please say your password after the beep.'' However, this message may be redundant for those who are used to it, and for those who are more accustomed to it, it may be better to omit the output of the signal sound.Especially in the case of inquiry-based services, calls are usually However, in the conventional system, redundant guidance and signal sound output are undesirable because they unnecessarily increase the holding time of the line and impose a burden on call charges.However, in conventional systems, There was only one type of response guidance, and it was impossible to change it according to the user's situation. Furthermore, although it is possible to prepare a plurality of guidances, the problem is how to decide which guidance should be output to which user. In other words, in conventional analog networks, since the receiving side does not know where the call is coming from, it is difficult to identify the user unless there is an opportunity for the user to enter the user's identification information manually. Providing such an opportunity was itself fraught with the contradiction that it would lead to an extension of the hold time. As described above, with conventional devices, inexperienced users tend to make mistakes in response procedures, and for experienced users, the sending of signal sounds and some of the guidance contents are redundant. Furthermore, in the case of Uokai-based services, the call fee is usually charged to the user who is the caller. Therefore, it is desired to eliminate wasteful guidance and shorten the length of time the line is held. The purpose of the present invention is to solve the problems of these conventional techniques, and to provide a detailed procedure for those who are not accustomed to operating it or those with poor recognition ability, while providing a simple procedure for those who are accustomed to operating it or those with good recognition ability. It is an object of the present invention to provide a response method of a voice response recognition device that can automatically select a suitable procedure and efficiently proceed with a service according to a wide range of users. [Means for Solving the Problems] In order to achieve the above object, the response method of the voice response recognition device of the present invention connects the voice response recognition device to a network having a caller telephone number notification function, and When a calling phone number is input, it recognizes the calling phone number, determines the operational proficiency level attached to this calling phone number, and selects a response guidance corresponding to the calling phone number from among multiple response guidance prepared in advance. The system is characterized in that it selects whether to send or not send a signal sound to prompt the user's voice input in accordance with the calling telephone number, and responds based on the selection result. [Function J] In the present invention, the voice response recognition device is provided with
A plurality of response guidances are prepared according to the user's level of proficiency in operating the system, and the system is connected to an ISDN network or the like. In ISDN networks and the like, the caller's telephone number can be notified to the other party as a network function. When the user manually enters the calling telephone number, the voice response recognition device determines the user's proficiency in operating the system based on this 7It call number. The operating proficiency level based on this telephone number is prepared in advance as a table within the device. Next, the voice response recognition device selects response guidance in accordance with the determined operating proficiency level, and also produces a signal sound to encourage the 111 user to easily perform voice input based on the procedure indicated by the response guidance. Select whether to send or not. Using this response method, the voice response recognition device can determine whether multiple response guidance or signal/question sending functions are necessary or not.
Since it is automatically selected according to the user, it becomes a flexible man-machine interface for the user. [Examples] Examples of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a voice response recognition device according to the present invention. It is composed of a network control section 1, a voice recognition section 2, a voice response section 3, and a main control section 4. The network control unit 1 is connected to the ISDN network. In this IsDN network, etc., the caller's telephone number can be notified to the other party as a network function. In addition to controlling the transmission function, the network control unit 1 also includes a platform a output unit 5 for performing recognition input at an appropriate timing,
It has a caller number receiving section 6 that receives the caller's telephone number notified from the network, and exchanges necessary information with the main control section 4. The speech recognition section 2 analyzes the input speech, compares it with standard speech patterns stored in advance, recognizes the input word, and reports the result to the main control section 4. The voice response section 3, under the support of the main control section 4,
Output voice guidance. The main control unit 4 not only controls each unit, but also performs business processes such as managing the progress of services and searching for stored data. In the ISDN network connected to the network control unit 1, the telephone number of the caller can be notified to the other party as a function of the network. Although this caller's telephone number does not represent the caller itself, in the case of an ordinary household, it can be considered as a fixed caller in a first-order approximation. Therefore,
When the system receives the caller's phone number, the system can determine the caller's familiarity with the system and select an appropriate response. With this configuration, this voice response recognition device can provide services politely to people who are not used to it or who have poor recognition performance, and efficiently to people who are used to it or have good recognition ability. , the necessity or necessity of multiple response guidance and signal sound transmission functions is automatically selected according to the user. FIG. 2 is a flowchart showing the operation of the voice response recognition device in FIG. This embodiment shows an example of a processing flow when a proficiency level table is used. Here, when a caller's telephone number is notified, correspondence is made between the telephone number and the user's level of familiarity with the system. That is, when a call from a user is input, it is first checked whether the caller's M-phone number is registered (step 201), and if it is not registered, the caller is assigned a proficiency level. If it is registered, the proficiency level table is searched using the phone number as a key, and the caller's proficiency level is referenced (step 202). For callers judged to be "fair", a process is selected to send a signal sound corresponding to their proficiency level (step 2).
03). For callers whose proficiency level is determined to be "expert" in step 202, normal guidance is immediately provided without transmitting the signal sound in step 203 (step 205). After selecting to send a signal sound in step 203, polite guidance (step 204) is provided to the caller whose proficiency level is determined to be "unskilled", and if the proficiency level is determined to be "average". For the caller who received the request, normal guidance (step 205) is provided. In this way, it is possible to automatically determine the caller's proficiency level and provide guidance that is optimal for the caller's proficiency level in an optimal manner. FIG. 3 is an explanatory diagram showing an example of a proficiency level table used in the proficiency level reference step in FIG. 2. The proficiency level 32 of each telephone user is shown corresponding to the caller's telephone number 3l. Here, the proficiency level is divided into three levels. These proficiency levels are determined by the system administrator based on historical information such as the number of times the caller (the user of the telephone with the corresponding number) has used the system and the recognition tendency to date. The proficiency level rQJ indicates a user who is unskilled in how to use the system, a proficiency level of "1" indicates a user who is proficient in exchange, and a user with a proficiency level of "2" indicates a skilled user. . FIG. 4 is an explanatory diagram showing a service provision for users whose proficiency level is rOJ in FIG. 3. The guidance voice/signal sound 4l sent from the voice response recognition device and the user's input voice 42 via the telephone respond to each other in a predetermined order, making it possible to perform the process desired by the user. . First, the voice response recognition device issues a message 43 saying, ``This is ** bank. If you agree, please say the service code one digit at a time following the signal tone.'' After this message, no voice is input from the user, and the message is unattended. After the message 43, the voice response recognition device emits a signal sound 45 called "beep", and the user who hears this signal sound receives the service code rlJ according to the content of the message 43.
Enter 46. When the service code NJ46 is input, the voice response recognition device then says, ``Balance inquiry.If you wish, please say ``Hi'' after the signal tone. ” message 47 is displayed. After the user confirms the signal sound 48 "Bit J",
Yes” and enter reply 49. In this way, detailed guidance is provided to less proficient users who are not familiar with how to operate the system. It should be noted that for users of average skill level, the same guidance as that provided by the conventional system shown in FIG. 6 is provided. FIG. 5 is an explanatory diagram showing service specifications for users whose proficiency level is "2" in FIG. 3. Only the guidance voice 51 is sent out from the voice response recognition device, and the signal sound is omitted. By responding to the guidance voice 51 with the user's input voice 42 via the telephone in a predetermined order, the process desired by the user becomes possible. First, the voice response recognition device issues a short message 53 saying, ``This is ** bank. Please provide the service code.'' The user immediately receives the service code roIJ5
4, the voice response recognition device 2 outputs a confirmation message 55 of the processing content corresponding to the service code roIJ54, saying, ``This is a balance inquiry.'' The user inputs the reply 56 as "yes". In this way, experienced users who are familiar with how to operate the system are provided with simple guidance, without cues and without redundant steps, allowing them to respond quickly. I can do it. The explanations shown in Figures 3 to 5 are just examples, and the system can be designed freely depending on the level of proficiency and the number of guidance depending on the type of service and the type of situation. . In addition, although we have described an example of selecting an appropriate service specification based on the user's proficiency level, the user
It is also possible to register service specifications in advance according to one's preferences and automatically select them based on the caller number. As described above, according to this embodiment, when a user's caller phone number is notified, an appropriate response guidance can be selected from a plurality of response guidances depending on the user's experience level, etc. It becomes possible to select the transmission of a signal sound, and it is possible to provide a man-machine interface that suits the user. [Effects of the Invention] According to the present invention, a careful procedure can be automatically performed for a person who is not accustomed to operating the device or a person with poor cognitive ability, and a simple procedure can be automatically performed for a person who is accustomed to operating the device or a person with a good cognitive ability. It is possible to efficiently provide voice response services to a wide range of users.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the voice response recognition device according to the present invention, FIG. 2 is a flowchart showing the operation of the voice response recognition device in FIG. 1, and FIG. 3 is a user's proficiency level. Explanatory diagram showing an example of a proficiency table for reference, No. 4
Figure 5 is an explanatory diagram showing the service specifications for users with a low proficiency level, Figure 5 is an explanatory diagram showing the service specifications for users with a high proficiency level, and Figure 6 is an explanatory diagram showing the service specifications for the voice input trigger of the conventional system. The explanatory diagram shown in FIG. 7 is a timing chart showing the relationship between the signal sound output of the conventional system and the operation of the voice recognition device. 1. Network control unit, 2: Voice recognition unit, 3, Voice response unit, 4:
Main control unit, 5: Signal sound sending unit, 6: Caller number receiving unit,
31: Caller phone number, 32. Proficiency level, 4l: Guidance voice/signal sound, 42. Input voice, 43: Message,
44: No input rQJ, 45: Signal sound, 46: Service code rlJ, 47: Message, 48: Signal sound, 49
: Return 1 yen, 51: Guidance voice, 52: Input voice, 5
3: Message, 54 Service code roIJ, 55
Message, 56, Reply, 61: Guidance voice/signal sound, 62/Input voice, 63. Message, 64 Signal sound, 65: Human power rOJ, 66 Signal sound, 67 Human power NJ
, 68: Message, 7l: Guidance, 72: Signal sound, 73: Voice input, 74: Voice recognition operation. 1-2 2-melted straw thatching 2nd figure 204 205 3rd figure 4th figure 6th figure

Claims (1)

[Claims] (1) A voice response recognition device that uses voice recognition technology that recognizes voices of unspecified speakers to identify and automatically respond to a process specified by a user by voice via a telephone, the voice response recognition device is connected to a network that has a calling phone number notification function, and when the calling phone number from the user is input, it recognizes the calling phone number and calculates the operating proficiency level that has been added to the calling phone number in advance. selects the response guidance corresponding to the calling telephone number from among a plurality of response guidances prepared in advance based on the operating proficiency level, and sends/rejects a signal sound to prompt the user to input voice. A response method for a voice response recognition device, characterized in that the voice response recognition device selects a telephone number corresponding to the calling telephone number, and responds based on the selection result.