JPH03220961A - Telephone voice reply device - Google Patents

Abstract

PURPOSE:To enhance the convenience of information input operation by revising the processing form to an input acoustic signal decided once properly even when identification discrimination to the input acoustic signal has an error so as to apply recognition identification processing to input information thereby returning a reply to the result of recognition identification processing. CONSTITUTION:A changeover section 4 drives selectively a voice recognition reply section 5 being a 1st reply means in recognition response to a voice input or a pushphone signal identification reply section 6 being a 2nd reply means in recognition response to a pushphone signal input according to the result of identification by a voice pushphone signal identification section 3. Then a voice reply message generated in response to the result of recognition identification of an input acoustic signal (voice or pushphone signal) by the reply sections 5, 6 is outputted as a reply from a voice output section 7 via an NCU (network control circuit) 1. Thus, the information is inputted while using the voice signal and the pushphone signal depending on the purpose.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a telephone voice response device that can effectively respond to voices and touch-tone signals input through a telephone line.

(Prior Art) Recently, various voice response services have been provided through telephone lines. For example, various telephone voice response service systems have been developed, such as balance inquiries and transfer inquiries in banking operations, and information on events at department stores and the like.

This type of system inputs specified information using a touch tone signal from the nearest telephone via a telephone line, or inputs specified information by telephone voice, and obtains a response to the input information by telephone voice. It is something. However, when the system detects an incoming call, it first identifies whether the acoustic signal input via the telephone line is voice or a touchtone signal, and then responds accordingly to the identification result. to switch and set the processing mode. If the input acoustic signal is identified as voice, the system instructs the user to input subsequent commands by voice, and while recognizing and processing the input voice, outputs a voice response message corresponding to the recognition result. This is then output as a response via the telephone line. In addition, if the input acoustic signal is identified as a touch tone signal, the system instructs the user to input subsequent commands using the touch tone signal, and responds to the identification result while processing the input touch tone signal. It is designed to generate and output a voice response message.

In other words, it identifies whether the acoustic signal input through the telephone line is a touch-tone signal or voice, selectively drives the voice recognition means or touch-tone signal identification means according to the identification result, and It provides a response service for signals (signals).

However, in this type of conventional system, the processing form of the subsequent voice response service is defined according to the identification result for the first input acoustic signal, and after that, the input signal is voice or only a touch-tone signal is input. The process will proceed under this premise. For this reason, if an error occurs in the identification processing for the first input acoustic signal, a problem arises in that response processing for subsequent command inputs becomes impossible. For example, if a silent period continues for a long time before a touch-tone signal is input, it may be mistakenly identified as voice.

In this case, since the voice recognition response section is selectively driven, there is a tendency to cause a problem in that it becomes impossible to identify touch-tone signals that are input thereafter.

Furthermore, when receiving a voice response service using a telephone in this manner, it is often possible to input commands more easily and efficiently by changing the information input format. For example, when inputting numerical information such as an account number, it is easier and more efficient to use a touch-tone signal, but when inputting commands for inquiries, etc., it is easier to use voice. be. However, as described above, if the information input form is defined as an alternative, a problem arises in that it is not possible to input information while selectively using touchtone signals and voice.

(Problems to be Solved by the Invention) As described above, in the conventional telephone voice response device (voice response service system), the input form of the information is inputted using a touch tone signal according to the identification result for the first input acoustic signal. However, since the system only specifies whether the information is input using voice or voice, the apparent performance is significantly lowered and the efficiency of information input is poor.

The present invention has been made in consideration of these circumstances, and its purpose is to provide an efficient voice response service by accurately recognizing and identifying voices or touchtone signals input through a telephone line. The object of the present invention is to provide a highly practical telephone voice response device that can be realized.

[Structure of the Invention (Means for Solving the Problems) The present invention identifies whether an acoustic signal transmitted through a telephone line is a voice or a touch-tone signal, and responds to the voice input according to the result of this discrimination. In a telephone voice response device that executes a voice response service by selectively driving a first response means or a second response means that makes an identification response in response to input of a touchtone signal, the first or second response means that is selectively driven. The present invention is characterized in that it is provided with means for changing the first and second response means that are selectively driven as described above in accordance with the voice recognition result or the push-tone signal identification result by the second response means.

That is, the present invention is characterized in that, when there is an error in the voice recognition result or the push-tone signal identification result by the first or second response means, means is provided for switching and controlling the response means for the input acoustic signal.

(Function) According to the present invention, even if there is an error in the identification result for the acoustic signal first input via the telephone line, the first or second response means is selectively driven according to the identification result. Since the response means is changed when an error occurs in the identification/recognition result, it is possible to reliably identify/recognize the input acoustic signal and perform appropriate response processing. Furthermore, even if the input form of information changes between voice and touch-tone signals, the response means can be switched accordingly. As a result, information can be input using voice and touch-tone signals.

(Embodiment) Hereinafter, a telephone voice response device according to an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a schematic configuration diagram of the embodiment device, and l is a network control unit (N
CU). The embodiment device is connected to the telephone line via this NCU1. Therefore, when an incoming call via the telephone line is detected at the NCUI, the acoustic signal input unit 2
It inputs and processes acoustic signals input through the telephone line.

The input acoustic signal is A/D converted by the acoustic signal input section 2 and is provided to the voice/touch-tone signal identification section 3 . The voice/push-telephone signal identification unit 3 is, for example, based on the following document "Proceedings of the 1981 Institute of Electronics and Communication Engineers Comprehensive National Conference "Push-touch phone signal and voice identification method using speech recognition device"
This system employs an identification method as introduced in "Page 6-217" to identify and determine whether the acoustic signal input through the telephone line is a voice or a touch-tone signal.

The switching section 4 has a voice recognition response section 5 which is a first response means that recognizes and responds to the voice input according to the identification result by the voice/touch-tone signal identification section 3, or a second voice recognition response section 5 that recognizes and responds to the touch-tone signal input. The touch-tone signal identification response section 6, which is a response means, is selectively driven. Then, a voice response message generated by these response units 5.6 according to the recognition/identification results of the input acoustic signals (speech or sound signals) is sent from the voice output unit 7 to the NC.
The response is output via the UI. still,
The voice recognition response unit 5, for example, as shown in FIG.
It recognizes and processes numeric sounds such as ``two'' and ``cue,'' as well as correct and incorrect voices such as ``hi'' and ``yay.'' Further, the touch-tone signal identification response section 6 detects the multiplexed frequency components of the touch-tone signal, and performs processing to identify numerical information and special symbol information indicated by the touch-tone signal. According to such recognition/identification results, a voice response message is generated according to the processing procedure shown in FIG. 3, for example, and is output as a response. To briefly explain the processing procedure of the voice response service in banking operations shown in Fig. 3, ■ When the bank is called from the nearest telephone, the center side detects this and receives a call) and says, ``This is OO Bank Telephone Service Center.'' Outputs a response message such as "Please enter your service code after the beep."

(2) Upon receiving such a response message, the user inputs predetermined service code information using a touch-tone signal, for example, by operating the dial pad of a telephone. In response to this, the center side identifies the input acoustic signal as a touch-tone signal, and the touch-tone signal identification response unit B
selectively drive. Then, the touch tone signal is identified and processed, and the following voice response messages are sent according to the service code that is the identification result: ``Balance inquiry.''``Today's current account deposit/withdrawal inquiry.'' This is an inquiry.'' etc. is output as a response. Then, it outputs voice messages one after another prompting information input, such as ``Course code please.'' ``Store number and account number please.'' ``Recitation number please.''

■■■ According to such a voice message, the user inputs the subject code using a touch-tone signal, and then inputs the store number, account number, and recitation number in order using the touch-tone signal.

The center side obtains this input information and outputs a voice message such as "Please wait.", then recognizes and processes this input information, and proceeds with a predetermined response processing procedure based on the recognition result. Then, after outputting a voice message saying "Thank you for waiting.", the response to that inquiry is output as a response voice message, followed by a message such as "Thank you. If you have any other requests, please send us your service code." is output, and if no service code is input, the voice response process ends at that point.

Such voice response processing can be carried out in the same way even when the input information (acoustic signal) is given in the form of voice, by selectively driving the voice recognition response section 5 instead of the touchtone signal identification response section 6.

Here, the feature of this embodiment device is that the switching change unit 8 monitors the recognition/identification results by the voice recognition response unit 5 and touch-tone signal recognition response unit 6, and detects errors in the recognition/identification results. In this case, it is provided with a function of changing and controlling the voice recognition response section 5 and touch-tone signal identification response section 6, which are used for recognition/identification response processing of input acoustic signals.

That is, in this embodiment, as shown in FIG. 2, the concept of the control processing is shown. In step A), the input acoustic signal is identified and judged, and in accordance with the identification result, the touch-tone signal identification processing is performed (step B), or the voice recognition processing is performed. (Step C)
Selectively operate the input acoustic signal to identify/recognize the input sound signal and check the identification/recognition results.If the identification/recognition results are incorrect, or if the identification/recognition processing cannot be performed (reject), the above-mentioned The present invention is characterized in that it has a control function that changes the identification/recognition response means for the input acoustic signal under the identification judgment processing performed.

Therefore, even if, for example, the input acoustic signal is identified as a touch-tone signal and a voice is input while the touch-tone signal identification response unit 6 is operating, the touch-tone signal identification response unit 6 In response to the recognition reject, the voice recognition response unit 5 is activated, and the input voice is recognized by the voice recognition response unit 5.

Conversely, if a touch-tone signal is input while the voice recognition response section 5 is operating, the touch-tone signal recognition response section 6 receives a recognition reject from the voice recognition response section 5.
is activated, and the touch-tone signal identification response section 6 processes the touch-tone signal for identification.

It goes without saying that if recognition/identification rejection is detected, the user is prompted to re-enter the information. Regarding recognition/identification rejects, for example, when a reject occurs for two consecutive inputs, it is determined that there is an error in the identification result for the input audio signal, and the response section is switched involuntarily. It is desirable to prevent this from happening.

Thus, according to the embodiment device having the above-mentioned functions, even if there is an error in the identification result for the first input audio signal, the response section is switched according to the recognition/identification result for the input audio signal. It becomes possible to correctly recognize the input acoustic signal and improve its apparent performance. Furthermore, even if voice and touch-tone signals are used to input information as appropriate, each of these pieces of information can be recognized, thereby improving information input processing efficiency and increasing the convenience of using the voice response system. A great practical effect can be achieved, such as:

Note that the present invention is not limited to the embodiments described above. For example, the voice recognition response section 5 and the P-noshphon signal identification response section 6 may be driven in parallel, and switching control may be performed to determine which recognition/identification result is to be adopted according to the recognition/identification results. It is also possible to control the switching using only one of the recognition/identification results of the voice recognition response section 5 and the touch-tone signal recognition response section 6. In addition, the present invention can be implemented with various modifications without departing from the gist thereof.

[Effects of the Invention] As explained above, according to the present invention, even if there is an error in the identification judgment for the input audio signal, the input information can be recognized by appropriately changing the processing form for the input audio signal whose amount has been determined. - Since it is possible to perform identification processing and return a response in response to the recognition/identification results, it has great practical effects, such as increasing the convenience of information input operations and efficiently proceeding with voice response processing.

[Brief explanation of drawings]

FIG. 1 is a schematic configuration diagram of a telephone voice response device according to an embodiment of the present invention, FIG. 2 is a diagram showing the concept of signal processing type switching control in the embodiment device, and FIG. 3 is an example of a voice response service. FIG. 4 is a diagram showing an example of the form of inputting information by voice. ■...Network control circuit (NCU), 2...Acoustic signal input unit, 3...Audio/touch-tone signal identification unit, 4...
- Switching unit, 5... Voice recognition response unit, 6... Touch-tone signal identification response unit, 7... Audio output unit, 8... Switching change unit.

Claims (1)

[Claims] means for inputting an acoustic signal via a telephone line; means for identifying whether the input acoustic signal is a voice or a touch-tone signal; a first response means for recognizing and responding to the voice input; The second
response means; means for selectively driving the first and second response means according to the identification results by the identification means; and voice recognition results and touch-tone signal identification by the first and second response means. 1. A telephone voice response device comprising: means for changing the first and second response means selectively driven according to the result.