JPH0133986B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a failure detection method for voice response devices such as those used in data telephone services and bank online services.

(Prior Art) Recently, there has been an increase in systems that use voice response devices to provide information requested by customers, such as in data telephone services and bank online services. Such a system can cause serious inconvenience to customers not only if the voice response device fails completely, but also if it misleads the customer with unclear voice output. A voice response device generally consists of a common control unit and a line-compatible voice synthesis output unit. Normally, a failure in the common control unit is easy to detect by the central control unit, but a failure in the line-compatible voice synthesis unit is difficult to detect.
A conventional method for detecting a failure in a speech synthesis unit detects a silent state in the output, and if the silent state continues for a certain period of time or more, it is regarded as a failure. This method has the disadvantage that it is not possible to detect disturbances that are not silent but have low message output clarity, and such disturbances can only be discovered through customer complaints.

(Objective of the Invention) An object of the present invention is to provide a voice synthesis device and a voice recognition unit, and to output test voice data from a voice response device, and to have the voice recognition device judge the voice data. An object of the present invention is to provide a voice response system that enables the detection of failures including a decrease in intelligibility and does not disturb customers.

(Structure of the Invention) A voice response system according to the present invention includes a voice response device, a voice recognition section, and a maintenance scanner display device (hereinafter referred to as "MSCN display device").

A voice response device has a plurality of voice synthesizers each connected to a plurality of lines accommodated in a call path switch via each trunk, and usually includes a line number, sentence pattern number, variable parameters, etc. The voice data is edited in the internal memory using the control information (hereinafter referred to as "CTL information") consisting of Holding the idle information or in-use information of the plurality of lines, applying a test activation to the idle lines, and having the voice synthesis unit corresponding to the idle line subjected to the test activation synthesize test voice data to generate voice. This is what outputs it.

The speech recognition section has speech recognition elements each connected to a plurality of lines and a flip-flop provided corresponding to the speech synthesis section, and when the test activation is applied, the flip-flop corresponding to the speech synthesis section is triggered. , a corresponding speech recognition element is for recognizing the test speech output.

The maintenance scanner display device is for displaying the recognition results of the voice recognition section as a maintenance scanner.

(Example) Next, the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment for realizing a failure detection method for a voice response device according to the present invention. In FIG. 1, 1-0 to 1-15 are speech synthesis units, and 2-0 to 2-15 are speech synthesis units 1-
Trunk for relaying outputs of 0 to 1-15,
3 is a call path switch, 4 is a voice data editing process during normal service, and a voice synthesis unit 1-0
~ 1-15 is an editing test control unit for performing test control; 5 is a control line for voice synthesis units 1-0 to 1-15;
6 is a voice recognition unit, 8 is a central control unit, and 7 is a display of the output of the voice recognition unit 6, which can be read in accordance with instructions from the central control unit 8.
The MSCN display device is a scanner that indicates whether the voice response output line (lines 0 to 15 in the example) is normal or abnormal, and is read by the central processing unit and used to determine whether the voice line is valid or invalid. 9 is used when the central controller 8 controls the editing test control section 4;
This is the bus used by the MSCN display device 7.
Further, 1 is a voice response device.

Figure 2 shows voice synthesis units 1-0 to 1- corresponding to the line.
15 is a diagram displaying the state of the internal memory of the edit test control section 4 used by the edit test control section 15. FIG.

FIG. 3 is a diagram showing the block configuration of the speech recognition unit 6. In FIG. 3, 10 is a flip-flop, 11-0 to 11-15 are speech recognition elements, and 12 is a speech recognition control section. The flip-flop 10 holds the number of the voice synthesis section 1-0 to 1-15 that is being tested. The speech recognition elements 11-0 to 11-15 correspond to the speech synthesis sections 1-0 to 1-15 and recognize their outputs. The speech recognition control unit 12 controls the speech recognition elements 11-0 to 11-15 and controls the MSCN display of the recognition results.

Normally, when sending a voice message to a customer, CTL information from the central control device 8 is sent to the editing test control section 4. The CTL information includes a line number, a sentence pattern number, a variable parameter, etc., and the editing test control unit 4 edits the voice data on the internal memory using the sentence pattern number and the variable parameter, and sends it to the designated voice synthesis unit. 1-i (if the i-th is specified).

The editing test control section 4 controls each speech synthesis section 1.
As shown in Figure 2, "Empty/Used" information to indicate whether or not -0 to 1-N is used by the above-mentioned normal services is maintained in memory, and if it is empty, it is Speech synthesis unit 1-j (j
(if applicable). The speech synthesis section 1-j that has been activated for the test outputs the test speech data held therein.
Since the speech recognition section 6 has a block configuration as shown in FIG. 3, when the speech synthesis section 1-j is started for testing, the corresponding flip-flop 10 is triggered by the editing test control section 4. The voice recognition control unit 12 of the voice recognition unit 6 controls the voice recognition element 11 corresponding to the triggered flip-flop 10.
-j to recognize the test audio output of the audio synthesizer 1-j. The voice recognition control unit 12 summarizes the recognition results of the voice recognition elements 11-0 to 11-15 and periodically sends them to the MSCN display device 7.
Note that the voice recognition elements 11-0 to 11-15
You may also experience problems yourself. The speech recognition elements 11-0 to 11-15 are provided as a pair, and when the output of the speech synthesis units 1-0 to 1-15 is recognized poorly, the speech recognition elements 11-0 to 11- 15 and perform the confirmation work again. The central control unit 8 can select a normal speech synthesis unit 1-0 to 1-15 by periodically reading the status information of the speech synthesis units 1-0 to 1-15 sent to the MSCN display device 7. It is possible.

(Effects of the Invention) As described above, the present invention includes a voice response device and a voice recognition unit, and transmits a test voice output from a corresponding voice synthesis unit of the voice response device to an idle voice line, By configuring this to be determined by a voice recognition device, it is possible to detect anything from a complete failure to a mild failure resulting in a decrease in intelligibility, which has the effect of improving service to customers. Furthermore, with the above configuration, the voice recognition unit can simultaneously check the idle voice lines in parallel, which has the effect of improving testing efficiency.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the block configuration of a voice response system according to the present invention.
FIG. 2 is a diagram showing an example of usage state display information of the speech synthesis section stored in the internal memory of the editing test control section in FIG. 1. FIG. 3 is a block diagram showing an embodiment of the detailed configuration of the speech recognition section in FIG. 1. 1-0 to 1-15...Speech synthesis section, 2-0 to 2
-15...Trunk, 3...Call path switch, 4
...Editing test control unit, 5...Control line, 6...Voice recognition unit, 7...MSCN display device, 8...Central control unit, 9...Bus, 1...Voice response device, 10
...Flip-flop, 11-0 to 11-15...
...Speech recognition element, 12...Speech recognition control unit.

Claims (1)

[Claims] 1 It has a plurality of speech synthesis units each connected to a plurality of lines housed in a call path switch via each trunk, and usually includes a line number, sentence pattern number, variable parameters, etc. The voice data is edited in the internal memory according to the information, and the synthesized voice synthesis output is sent to the line by the voice synthesizer corresponding to the specified line among the plurality of lines, and the information on the availability or use of the plurality of lines is sent. a voice response device that holds a test activation unit for a vacant line, causes a voice synthesizer corresponding to the idle line subjected to the test activation to synthesize test voice data and output voice; and the plurality of lines. and a flip-flop provided corresponding to the speech synthesis section, respectively, and when the test activation is applied, the flip-flop corresponding to the speech synthesis section is triggered, and the corresponding speech recognition element is activated. 1. A voice response system comprising: a voice recognition section that recognizes a test voice output; and a maintenance scanner display device that displays the recognition results of the voice recognition section on a maintenance scanner.