JP5624451B2 - Calling device and calling method - Google Patents

Description

  The present invention relates to a calling device and a calling method.

  Call centers that conduct reminders, sales, surveys, announcements, etc. are automating outbound and inbound services using CTI (Computer telephony integration) technology to reduce costs. Some conventional call systems that perform outbound services automatically make a call and connect to an operator by a response from the other party (see, for example, Patent Document 1).

JP 2004-266598 A

At least when using the Japanese telephone network, it is possible to identify whether the callee answers, is busy, or is not answered by the call system.
However, even if a person responds on the called side, or if the answering function of the telephone answering machine or an announcement function that notifies that it is out of service, the above call system can recognize only as the same response. All are connected to the operator, and the operator listens to the sound and determines whether the response is a human response or a machine response. This adversely affects the rate of reaching the conversation with the called party.

  In particular, in the case of an outbound service used in a dunning service, a called party often intentionally switches to an answering machine function, which can require a large amount of cost for identifying the answering machine function.

  The present invention has been made in view of these problems, and an object of the present invention is to provide a technique capable of more appropriately discriminating the subject of a response on the called side in an outbound system that automatically makes a call.

  One embodiment of the present invention relates to a calling device. This calling device extracts a telephone number to be called from a list of telephone numbers from a line control unit that controls communication with a communication device via a telephone network, and makes a call to the telephone number to the line control unit. And a determination unit that determines whether or not a response from the communication device on the called side is due to a person by analyzing a signal acquired from the communication device on the called side by the line control unit When the determination unit determines that the response is not a human response, the line control unit monitors the signal acquired from the called communication device and indicates that recording is started at the called communication device. A standby unit that waits for a signal corresponding to the sound.

  According to this aspect, the determination result of whether or not the response from the communication device on the called side is due to humans and the detection of the signal corresponding to the sound indicating that the recording on the called communication device is started Based on the presence or absence, the subject of the response on the called side can be determined.

  It should be noted that any combination of the above-described constituent elements, or those obtained by replacing the constituent elements and expressions of the present invention with each other between apparatuses, methods, systems, computer programs, recording media storing computer programs, and the like are also included in the present invention. It is effective as an embodiment of

  According to the present invention, in the outbound system that automatically makes a call, the subject of the response on the called side can be more appropriately determined.

It is a schematic diagram which shows the structure of a telephone system. It is a block diagram which shows the function and structure of the calling device which concerns on 1st Embodiment. It is a data structure figure which shows an example of the list holding | maintenance part of FIG. It is a data structure figure which shows an example of the message holding | maintenance part of FIG. It is a flowchart which shows an example of a series of processes in the calling device of FIG. 6 is a flowchart for explaining in more detail a step of waiting for a “P sound” in the flowchart shown in FIG. 5. 7 (a), (b), and (c), respectively, when a response by the answering machine function is returned from the called side, when a person responds at the called side, and from the called side, etc. It is a figure which shows the time change of a typical frequency spectrum when the announcement of the carrier for this is returned. FIGS. 8A and 8B are diagrams illustrating typical frequency spectra of signals corresponding to “P sound” and sound, respectively. It is a block diagram which shows the function and structure of the calling device which concerns on 2nd Embodiment. It is a flowchart which shows an example of a series of processes in the calling device of FIG.

  A solution is also conceivable in which a line control signal is used to determine whether a response is made by a human or an answering machine function. However, the present inventors have recognized that this is difficult in the following points.

There are at least two types of answering machine functions in Japan: the answering machine function provided by the carrier of the telephone network with which the called terminal is contracted, and the answering machine function possessed by the called terminal itself.
Hereinafter, problems in the case of using a line control signal will be described.

(1) Voice mail service provided by the carrier of the telephone network When the telephone line on the calling side is an analog line and the answering machine function of the carrier responds, only the loop signal is returned to the calling side line and an announcement flows. When the announcement ends, a reverse signal is returned and the caller can leave a message (during a call).

  When the caller's telephone line is a digital (ISDN (Integrated Service Digital Network)) line and the answering service of the carrier responds, the progress indicator (PI (Progress Indicator)) recognizes that an announcement will flow on the caller it can. The caller can leave a message (during a call) by a response (CONNECT) message to end the announcement.

  As described above, in the answering service provided by the carrier, it is possible to grasp the answering answering announcement of the answering machine and the recording timing after the announcement is finished by the line control signal.

However, the connection control of the line control signal is not notified as it is when the carrier of the telephone network on the calling side is different from the carrier on the telephone network on the calling side, or when the carrier of the telephone network passing through is different from the calling side or the called side. For this reason, there may be a case where the telephone is in a busy state from the time when the announcement of the answering machine function starts to flow or the progress indicator becomes unavailable.
Therefore, it is difficult to identify all responses of the answering machine function provided by the carrier of the telephone network only by the information of the line control signal.

(2) Answering machine function of telephone terminal When answering by the answering machine function of the telephone terminal, the same line control signal as the response of the called terminal is returned to the calling telephone line. It is difficult to determine from the line control signal whether or not the telephone terminal has responded or the answering machine function of the telephone terminal has responded.

As described above, when depending only on the line control signal, a situation may occur in which it is difficult to determine whether the response is based on the answering machine function.
For this reason, there is a method of recognizing the announcement of the answering machine function of the carrier or the guidance of the answering machine function of the telephone terminal using voice recognition without relying on the line control signal. However, it is difficult to respond when the user changes the response message of the (A) answering machine function, and (B) it is difficult to grasp all the response guidance that is different for each telephone terminal in terms of speech recognition. For this reason, it is difficult to accurately identify the response by the answering machine function even if voice recognition is used.

  In general, at least in Japan, if the answering machine function responds to whether it is a telephone terminal or a carrier, the answering machine function will start recording at the called terminal. A beep sound (hereinafter referred to as “P sound”), in other words, a sound for giving the caller the timing to start speaking the recorded message or prompting the utterance of the recorded message is played. The present inventors examined whether it is not possible to determine whether or not the answering by the answering machine function is detected by detecting the “P sound”.

  In general, the signal corresponding to the “P sound” is a simple single sine wave and further flows for a short period of time of a few hundred ms. Therefore, when relying only on the detection of the “P sound”, particularly in the method of detecting the spectral component of the “P sound”, the sound component of the “P sound” in the sound of the response message and the environmental sound when a human responds. May be erroneously detected. Therefore, although it is possible in principle to identify the answering machine function based only on the “P sound”, it is desired to further refine it in practice.

  The present invention will be described below based on preferred embodiments with reference to the drawings. The same or equivalent components, members, and processes shown in the drawings are denoted by the same reference numerals, and repeated descriptions are omitted as appropriate.

(First embodiment)
FIG. 1 is a schematic diagram showing the configuration of the telephone system 2. The telephone system 2 includes a calling device 100 according to the first embodiment, a called telephone terminal 4, a telephone network 6 such as a PSTN (Public Switched Telephone Network), a private branch exchange (PBX (Private Branch eXchange)) 8 The operator terminal 10 is provided. The calling device 100, the private branch exchange 8 and the operator terminal 10 are provided in a call center 12. The calling device 100 and the operator terminal 10 are each connected to the private branch exchange 8. Private branch exchange 8 is connected to telephone network 6. The telephone terminal 4 is connected to the telephone network 6.

  In the calling device 100 according to the present embodiment, the following ideas (A) and (B) are combined to determine whether or not the response is made by the answering machine function with higher accuracy.

  (A) When the called party answers the outgoing call, the calling device 100 determines whether or not the response from the called party is from a human. An example of this determination method is as follows. When the called party responds to the call, the calling device 100 sends a message (hereinafter referred to as an initial guidance message) including the company name and purpose of the calling party to the called side. At this time, the calling device 100 determines that the response is a machine response when it continuously receives a voice of a certain time or more from the called party. Whether or not the sound being played from the called party is a voice is identified by performing a spectrum analysis of the sound and indicating whether or not the spectrum represents the voice band.

  At this point, even though it can be identified that the response is by a machine, it cannot be identified whether it is a response by an answering machine function or an announcement from a carrier of the telephone network 6 for other reasons.

  (B) In response to the answering machine function, whether it is a telephone terminal 4 or a carrier, a response message (hereinafter referred to as an answering machine response message) for notifying that the called party is answering or cannot answer the call. "P sound" flows after. Therefore, when it is determined that the response is not from a human, the calling device 100 waits for a “P sound”. When detecting the “P sound” while waiting for the “P sound”, the calling device 100 determines that the response on the called side is a response by the answering machine function.

In general, since the “P sound” is a simple single sine wave, the calling device 100 determines the presence or absence of other spectral components when detecting this spectral component. Then, the calling device 100 detects the “P sound” on condition that a frequency spectrum having a peak only at the frequency corresponding to the “P sound” continues for a predetermined time. When this condition is satisfied, the calling device 100 executes a process to be executed when the “P sound” is detected, that is, sends a message to be recorded by the called party (hereinafter referred to as a recording message). To start.
By determining the presence or absence of other spectral components, the possibility of erroneous detection of “P sound” is reduced.

  FIG. 2 is a block diagram showing the function and configuration of calling device 100 according to the present embodiment. Each block shown here can be realized by hardware such as a computer (CPU) (central processing unit) and other elements and mechanical devices, and software can be realized by a computer program or the like. Here, The functional block realized by those cooperation is drawn. Therefore, it is understood by those skilled in the art who have touched this specification that these functional blocks can be realized in various forms by a combination of hardware and software.

  The calling device 100 includes a line control unit 102, a call control unit 104, a first determination unit 106, a second determination unit 108, a transfer control unit 110, a standby unit 112, a recorded message transmission control unit 114, a list holding unit 116, A message holding unit 118.

  FIG. 3 is a data structure diagram illustrating an example of the list holding unit 116. The list holding unit 116 holds a list of telephone numbers of persons who are call candidates. In particular, the list holding unit 116 holds the name of a person who is a call candidate, a telephone number, and a plurality of attributes in association with each other. The attribute is, for example, the sex or age of a person who is a call candidate.

  FIG. 4 is a data structure diagram illustrating an example of the message holding unit 118. The message holding unit 118 holds a message ID for specifying a message to be sent from the calling device 100 to the called party, a message type, and a message content in association with each other. Examples of the message type include an initial guidance message, a recorded message, a response control question, and an expected response for response control.

  The line control unit 102 is connected to the private branch exchange 8 by wire or wirelessly. The line control unit 102 is connected to the telephone network 6 via the private branch exchange 8. The line control unit 102 controls communication with a communication device such as a fixed telephone terminal, a mobile phone terminal, or a carrier communication control device in the telephone network 6 via the telephone network 6. The line control unit 102 may be configured based on various communication standards.

The call control unit 104 controls a call from the call device 100. The call control unit 104 includes a call target extraction unit 120 and a call instruction unit 122.
The call target extraction unit 120 sequentially selects telephone numbers from the list holding unit 116 and extracts them as call target telephone numbers.

  The call instruction unit 122 causes the line control unit 102 to call the telephone number extracted by the call target extraction unit 120. The call instruction unit 122 passes the telephone number extracted by the call target extraction unit 120 to the line control unit 102. When the line control unit 102 acquires a telephone number from the call instruction unit 122, the line control unit 102 makes a call to the telephone number.

  The first determination unit 106 determines whether the communication device on the called side has responded, is busy, or is not responding in response to the call addressed to the telephone number extracted by the call target extraction unit 120. . The first determination unit 106 determines that the communication device on the called side has responded when the line enters a busy state corresponding to the outgoing call.

  In addition, the first determination unit 106 acquires from the line control unit 102 a line control signal of a line used for a call addressed to the telephone number extracted by the call target extraction unit 120. The first determination unit 106 determines whether or not an announcement flows from the carrier of the telephone network 6 with reference to the line control signal. For example, the first determination unit 106 refers to the progress indicator information element included in the line control signal.

  In addition, the first determination unit 106 is configured so that the line control unit 102 receives the call from the communication device on the called side even when the progress indicator information element is not available and the line is not busy in response to the outgoing call. When a signal corresponding to voice is detected in the acquired signal, it is determined that an announcement from the carrier is flowing. The signal acquired by the line control unit 102 from the called communication device may be an electrical signal acquired by the line control unit 102 from the line connecting the line control unit 102 and the called communication device.

When the first determination unit 106 determines that the called communication device has responded and the second determination unit 108 determines that the announcement from the carrier does not flow, the second control unit 108 determines that the line control unit 102 By analyzing the signal acquired from the communication device, it is determined whether or not the response from the communication device on the called side is from a human.
Note that even when the first determination unit 106 determines that no announcement from the carrier flows, the announcement from the carrier may actually flow for the reason described in (1) above.

The second determination unit 108 is configured to be able to use any one of a guidance-continuous voice detection mode, a response control mode, and an emotion analysis mode described below, or any combination thereof.
The second determination unit 108 includes an initial guidance transmission control unit 124, a continuous voice detection unit 126, a reception control unit 128, an emotion analysis unit 130, and a change amount determination unit 132.

(Guidance-Continuous voice detection mode)
In the guidance-continuous voice detection mode, the second determination unit 108 causes the line control unit 102 to send a signal corresponding to the initial guidance message to the communication device on the called side. When the second determination unit 108 returns a continuous voice of a predetermined length or more from the called communication device after transmission of a signal corresponding to the initial guidance message is started, the second determining unit 108 It is determined that the response from is not by humans.

The initial guidance transmission control unit 124 determines that the caller side communication device has responded in the first determination unit 106, and determines that the announcement from the carrier does not flow. Extract the content of the guidance message. In particular, the initial guidance transmission control unit 124 extracts the message content corresponding to the message type “initial guidance message” from the message holding unit 118.
The initial guidance transmission control unit 124 instructs the line control unit 102 to transmit a signal corresponding to the content of the extracted initial guidance message to the communication device on the called side.

  The continuous voice detection unit 126 monitors a signal that the line control unit 102 acquires from the communication device on the called side after transmission of a signal corresponding to the initial guidance message is started. When the continuous voice detection unit 126 continuously detects a signal corresponding to the voice among the signals being monitored for a predetermined continuous voice detection time or more, the response from the communication device on the called side is not caused by a human being. Determine. This is because if the response from the communication device on the called side is by a person, it will usually be silent when the initial guidance message starts to flow. However, when the initial guidance message starts to flow, there is a possibility that the called party may say “Hello” or say his name, which may overlap the initial guidance message. 126 may monitor a signal that the line control unit 102 acquires from the communication device on the called side starting from the end of the signal corresponding to the initial guidance message. In this case, the accuracy of determination can be improved.

  The continuous voice detection unit 126 performs frequency analysis on a signal acquired by the line control unit 102 from the communication device on the called side, and acquires a frequency spectrum. When the frequency spectrum indicating the spectrum of the voice band continues for the continuous voice detection time or longer, the continuous voice detecting unit 126 determines that the response from the communication device on the called side is not due to a human being.

  The continuous voice detection unit 126 only needs to be able to determine whether or not there is a voice in the response from the called communication device after the transmission of the signal corresponding to the initial guidance message is started. For example, it is set shorter than the time required for uttering a general number of characters.

  The continuous voice detection unit 126 determines that the response from the communication device on the called side is a human if the length of the period during which no signal corresponding to the voice is detected in the monitored signal exceeds a predetermined silent duration. It is determined that If the duration of silence is too short, the possibility of misjudgment increases, and if it is too long, if the called party is a human, it will wait for a long time for discrimination. Good.

(Response control mode)
In the response control mode, the second determination unit 108 determines whether or not the response from the communication device on the called side is due to a human by using IVR (Interactive Voice Response) technology.
When it is determined by the first determination unit 106 that the called communication device has responded and it is determined that the announcement from the carrier does not flow, the reception control unit 128 notifies the line control unit 102 to the called side. A response from the called communication device by sending a signal corresponding to the question to the called communication device and comparing the response from the called communication device to the expected response to the question It is determined whether or not is due to human beings. The response control unit 128 extracts a question and an expected response from the message holding unit 118 to the called party. The response control unit 128 determines that the response from the incoming communication device is not a human response when the response from the incoming communication device differs from the expected response by more than a predetermined degree.

(Emotion analysis mode)
In the emotion analysis mode, the second determination unit 108 performs emotion analysis on the response from the called communication device, and if the change in emotion is within a predetermined range, the response from the called communication device is received by a human. It is determined that it is not.

  Among the answering machine response messages of the telephone terminal 4, the default message and the announcement from the carrier are often mechanically synthesized voices. There is little emotional change in this synthesized speech. Further, if the voice file used for the answering machine response message is mechanically recorded, it is considered that there is little emotional change in the answering machine response message.

  Regarding emotion analysis used in the second discriminating unit 108, the speech includes broadly phonological information and prosodic information. In general, prosodic information is used as a parameter for specifying emotion from speech. This is because the main element that characterizes emotion is thought to be prosodic information. Emotion determination is performed by analyzing feature values of prosodic information. The feature amount is, for example, an amount that characterizes the inflection of speech, an amount that characterizes the size of speech, or an amount that characterizes the frequency of speech.

  If it is determined by the first determination unit 106 that the call side communication device has responded and the emotion analysis unit 130 determines that the announcement from the carrier does not flow, the line control unit 102 determines that the call side communication device A signal acquired from the apparatus is analyzed in real time to calculate a feature amount of prosodic information.

  When the state in which the feature amount calculated by the emotion analysis unit 130 is within a predetermined upper limit value and lower limit value continues for a predetermined emotion determination time, the change amount determination unit 132 performs communication on the called side. Determine that the response from the device is not human. When the calculated feature amount exceeds the upper limit value or the lower limit value, the change amount determination unit 132 determines that the response from the communication device on the called side is from a human.

  In consideration of the fact that the default value of the feature value of the prosodic information varies depending on the type of the answering machine, the second determination unit 108 first calculates the default value of the feature value of the prosodic information from the input speech, Analysis and discrimination may be performed as a reference. That is, the second discriminating unit 108 extracts a part of a predetermined length, for example, a few seconds from the signal acquired by the line control unit 102 from the called communication device, and the default value of the feature value of the prosodic information from the extracted part May be calculated.

  If the second determination unit 108 determines that the response from the communication device on the called side is from a human, the transfer control unit 110 may cause the line control unit 102 to transfer the call to the operator terminal 10. Alternatively, automatic response control on the premise that the called party is a human may be performed by the response control unit 128, for example.

  The standby unit 112 determines that the second determination unit 108 determines that the response from the communication device on the called side is not a human response, or the first determination unit 106 determines that an announcement is flowing or is flowing from the carrier. If so, the line control unit 102 monitors a signal acquired from the communication device on the called side, and waits for a signal corresponding to “P sound”. In particular, the standby unit 112 has a frequency spectrum having a peak only at a predetermined P sound corresponding frequency for a frequency spectrum of a signal acquired by the line control unit 102 from the communication device on the called side, for a predetermined P sound duration or longer. As a condition, a signal corresponding to “P sound” is detected.

The frequency corresponding to the P sound and the P sound duration are set in advance based on the “P sound” of the answering machine function of each carrier and the “P sound” of the answering machine function of various telephone terminals.
The standby unit 112 includes a spectrum acquisition unit 134, a level determination unit 136, and a continuation determination unit 138.

  The spectrum acquisition unit 134 determines that the second determination unit 108 determines that the response from the communication device on the called side is not caused by a person, or the first determination unit 106 receives or flows an announcement from the carrier. When the determination is made, the line control unit 102 performs frequency analysis on the signal acquired from the called communication device and acquires the frequency spectrum.

  In the frequency spectrum acquired by the spectrum acquisition unit 134, the level determination unit 136, when the component near the P sound corresponding frequency is a predetermined value or more and the other components are another predetermined value or less, the frequency The spectrum is determined to be a frequency spectrum that can correspond to “P sound”. The level determination unit 136 repeats the above process every time the spectrum acquisition unit 134 acquires a frequency spectrum.

  The continuation determination unit 138 detects the signal corresponding to the “P sound” when the level determination unit 136 continuously determines that the frequency spectrum can correspond to the “P sound” for the P sound duration or longer. It is determined that Further, the continuation determination unit 138 responds to “P sound” when it is not determined that a signal corresponding to “P sound” has been detected even after a predetermined P sound detection time has elapsed from the start of frequency analysis by the spectrum acquisition unit 134. It is determined that no signal has been detected.

  When it is determined that the signal corresponding to the “P sound” is not detected in the standby unit 112, the transfer control unit 110 may cause the line control unit 102 to transfer the call to the operator terminal 10. Alternatively, the standby unit 112 may cause the line control unit 102 to disconnect the call.

When the standby unit 112 detects a signal corresponding to “P sound”, the recorded message transmission control unit 114 extracts the content of the recorded message to be sent from the message holding unit 118. In particular, the recorded message transmission control unit 114 extracts the message content corresponding to the message type “recorded message” from the message holding unit 118.
The recording message transmission control unit 114 instructs the line control unit 102 to transmit a signal corresponding to the content of the extracted recording message to the called communication device.

The operation of the calling device 100 having the above configuration will be described.
FIG. 5 is a flowchart illustrating an example of a series of processes in the calling device 100. The calling device 100 automatically makes a call to a telephone number included in the telephone number list held in the list holding unit 116 (S202). The calling device 100 determines whether there is an announcement from the carrier (S204). When it is determined that there is an announcement from the carrier (Y in S204), the calling device 100 waits for “P sound” (S210). When it is determined that there is no announcement from the carrier (N in S204), the calling device 100 determines whether there is a response from the communication device on the called side (S206). When it is determined that there is no response from the communication device on the called side (N in S206), the calling device 100 disconnects the call (S214). When it is determined that there is a response from the called communication device (Y in S206), the calling device 100 determines whether or not the response from the called communication device is human (S208). ). When it is determined that the response is not from a human (N in S208), the calling device 100 waits for a “P sound” (S210). When the “P sound” is not detected (N in S210), the calling device 100 disconnects the call (S214). When the “P sound” is detected (Y in S210), the calling device 100 sends a recording message to the communication device on the called side (S212). When the sending of the recording message is completed, the calling device 100 disconnects the call (S214).

If it is determined in step S208 that the response is from a human (Y in S208), the calling device 100 performs appropriate automatic answering control (S216). Thereafter, the calling device 100 disconnects the call (S214).
The calling device 100 repeats the above-described series of processes for other telephone numbers included in the telephone number list held in the list holding unit 116.

  FIG. 6 is a flowchart for explaining in more detail the step S210 of waiting for the “P sound” in the flowchart shown in FIG. The calling device 100 performs frequency analysis on the signal acquired by the line control unit 102 from the called communication device, and acquires a frequency spectrum (S218). The calling device 100 determines whether or not the acquired frequency spectrum has a peak only at the frequency corresponding to the P sound (S220). When it is determined that only the P sound corresponding frequency has a peak (Y in S220), the calling device 100 determines whether or not such a frequency spectrum continues for the P sound duration (S222). Otherwise (N in S220), the calling device 100 acquires a frequency spectrum again (S218). When the P sound duration has continued (Y in S222), the calling device 100 determines that the “P sound” has been detected (S224). If the P sound duration has not been continued (N in S222), the calling device 100 determines whether or not the P sound detection time has elapsed since the start of the step of waiting for the “P sound” (S226). When the time has not elapsed (N in S226), the calling device 100 acquires a frequency spectrum again (S218). If it has elapsed (Y in S226), the calling device 100 determines that “P sound” has not been detected (S228).

  In the embodiment described above, examples of the holding unit are a hard disk and a semiconductor memory. Further, based on the description of the present specification, each unit is configured by a CPU (not shown), a module of an installed application program, a module of a system program, a semiconductor memory that temporarily stores the content of data read from the hard disk, or the like. It will be understood by those skilled in the art who have touched this specification that it can be realized.

FIGS. 7A, 7B and 7C respectively show a case where a response by the answering machine function is returned from the communication device on the called side, a case where a person responds on the called side, and a communication on the called side. It is a figure which shows the time change of a typical frequency spectrum when the announcement of the carrier for telling out-of-service etc. is returned from the apparatus.
In each of FIGS. 7A, 7B, and 7C, the horizontal axis represents time, and the vertical axis represents frequency. The intensity of the frequency component of the signal from the communication device on the called side is represented by black and white shading on a gray background, and the gray background becomes zero and gradually increases from dark to light. It represents the strongest thing.

  FIG. 7A is a diagram showing a time change of a typical frequency spectrum when a response by the answering machine function is returned from the communication device on the called side. When the answering machine function responds, the answering machine response message 302 flows without depending on the situation of the calling party, and finally the “P sound” 304 flows. Although the contents of the answering machine response message 302 are various, since the “P sound” 304 is a common sinusoidal wave, a spectrum in which high intensity (peak) appears only at a specific frequency, that is, near the P sound corresponding frequency 306, Obviously it is different from voice.

  FIG. 7B is a diagram showing a time change of a typical frequency spectrum when a person responds on the called side. When a person responds, the spectrum 308 appears a little because he / she says his / her name first, but since the initial guidance message is sent from the calling device 100, it is understood that the person who responded is a call from the machine. Because of silence, silence continues.

  FIG. 7C is a diagram showing a time change of a typical frequency spectrum when a carrier announcement for transmitting out-of-service or the like is returned from the communication device on the called side. There are various types of announcements 310, such as when a call is made to a mobile terminal outside the service area, or when a call is made to an unused telephone number, but this is different from the answering machine response. The point is that there is no “P sound”.

FIGS. 8A and 8B are diagrams illustrating typical frequency spectra of signals corresponding to “P sound” and sound, respectively.
FIG. 8A is a diagram showing a typical frequency spectrum of a signal corresponding to “P sound”. When the frequency spectrum analysis of “P sound” is performed, a level appears only at a specific frequency (frequency corresponding to P sound) n (Hz) as shown in FIG. 8A, and the level of other frequencies does not increase. . And such a state continues for a fixed time. In the present embodiment, the “P sound” is detected using such a property of the “P sound”.

  FIG. 8B is a diagram showing a typical frequency spectrum of a signal corresponding to voice. When frequency spectrum analysis is performed on a human voice or voice generated by machine synthesis, a wide range of frequency components appears as shown in FIG. The level changes with time, but since the level always appears in a wide range of frequency components, it can be identified as the “P sound”. Another feature is that a certain silent section appears between words and phrases, and the end of the talk can be identified by detecting silence for a certain period of time.

  According to the calling device 100 according to the present embodiment, when an automatic call is made, it is possible to more accurately determine whether a response is made by an answering machine function or not. Here, it does not matter whether the answering machine function is for the telephone terminal on the called side or for the carrier. Thereby, for example, when the calling device 100 is installed in a call center, a more efficient outbound service can be realized.

  The answering machine response message has various contents, inflections, tone, length, and the like, depending on the carrier, the telephone terminal, and the user. However, the “P sound” that tells the start of recording generally has a common property. In this embodiment, paying attention to the commonality of the “P sound”, it is determined whether or not it is an answering machine function. In addition, since the property of “P sound” is simple, there is a possibility that erroneous detection may occur if only it is relied on. Therefore, in the present embodiment, the possibility of such erroneous detection is reduced by providing a step for determining whether or not the response is made by a human in the previous stage.

  Further, when the calling device 100 according to the present embodiment determines that the response is based on the answering machine function, the calling device 100 automatically sends the recorded message to the communication device on the called side. As a result, the outbound service realized by the calling device 100 is more highly automated, reducing the burden on the operator and reducing the cost.

  Further, calling device 100 according to the present embodiment has an emotion analysis mode in which emotion analysis is applied to discriminating whether a response is made by a human or not. Here, the determination is made based on the change in emotion and the amplitude, and the type of emotion is irrelevant. Therefore, sufficient discrimination accuracy can be obtained without using such an advanced emotion analysis engine.

  In addition, the calling device 100 according to the present embodiment has a reception control mode in which reception control is applied to determine whether a response is made by a human or not. Thereby, the accuracy of discrimination is improved. It can also handle recognition of free speech.

(First application)
A first application that applies the calling device 100 according to the first embodiment to a medicine reminder service will be described. In the drug reminder business, the pharmaceutical company tracks the customer's remaining amount of drug and places a phone call to recommend additional purchases to customers who may be low. In this application, the list holding unit 116 holds a list of customer telephone numbers that should be recommended to purchase additional medicines.

  When the guidance-continuous voice detection mode is used, the message holding unit 118 holds the message content “I'm YY of XX Pharmaceutical Company. Are you ZZ?” For the message type “initial guidance message”. . The last name of the customer that can be extracted from the list holding unit 116 is inserted into ZZ. The message holding unit 118 indicates that the message type “recorded message” indicates the purpose of the call “I am always indebted. My name is YY of the XX pharmaceutical company. I understand that the number of tablets for WW's medicine for WW has been reduced. I will contact you again. "

  The calling device 100 calls a certain customer's telephone number extracted from the list holding unit 116. When the called party answers, the calling device 100 sends an initial guidance message, “What if you say YY of XX pharmaceutical company? Are you ZZ?” The calling device 100 determines that the response is not a human response when voice is continuously detected within a period of several tens of milliseconds after the initial guidance message starts to flow. If the calling device 100 determines that the response is not a human response and further detects a “P sound”, “I am always indebted. My name is YY from the XX pharmaceutical company. I'm calling you, and I'm sure there are fewer ZZ medicine tablets for WW. We will contact you again. " This leaves a message telling the purpose of the call to the answering machine function of the customer's telephone terminal or to the answering machine function of the carrier with which the customer is contracted.

When using the answering control mode, the message holding unit 118 holds the following question to the called party and the expected response thereto.
Question 1 to the callee: “Hello, my name is YY of the XX pharmaceutical company. Are you ZZ?”
Expected response 1: “Yes.
Question 2 to the callee: “I am always indebted. I'm calling you today with a guide to medicines. I know that ZZ ’s medicine tablets for WW have been reduced in number. ”
Expected response 2: “Oh, that was right”
Question 3 to the callee: “M months, I will arrange it again.
Expected response 3: “Yes, please”
Callee's Question 4: “Thank you. I will send you tomorrow. The total amount will be N yen. I will enclose the invoice. Thank you for your cooperation.” (Methods, deposit methods, etc. vary depending on the service model, so prepare the expected grammar in advance.)

  The calling device 100 determines that the response is not a human response when the response from the incoming call side deviates from the above-described response flow. If the calling device 100 further detects a “P sound” after determining that it is not a human response, the calling device 100 plays a recording message.

(Second application)
A second application that applies the calling device 100 according to the first embodiment to an emergency response announcement work such as an emergency earthquake warning will be described. In the emergency response announcement business, when an emergency occurs, calls are made to the telephone numbers registered in the system all at once to notify the emergency. In this application, the list holding unit 116 holds a list of telephone numbers of persons to be announced.

  When using the guidance-continuous voice detection mode, the message holding unit 118 responds to the message type “initial guidance message” with the message content “This is the XX Meteorological Agency. Currently, the YY river reaches the dangerous water area near your home. "Please condemn the emergency evacuation shelter (high altitude) near you as it may break down." The message holding unit 118 holds the same message content as the initial guidance message for the message type “recorded message”.

  The calling device 100 makes a call to a plurality of telephone numbers extracted from the list holding unit 116 simultaneously. When the communication device on the called side answers, the calling device 100 is “This is the XX Meteorological Agency. Currently, the YY river has reached the dangerous water area near your home. Please send an initial guidance message saying "Please condemn the shelter (high place)." The calling device 100 determines that the response is not a human response when voice is continuously detected within a period of several tens of milliseconds after the initial guidance message starts to flow. If the calling device 100 detects a “P sound” after determining that it is not a human response, “This is the XX Meteorological Agency. Currently, the YY river has reached a dangerous water area near your home. "Please accuse me of emergency shelters (high places) near you." Furthermore, the calling device 100 registers the telephone number that has played the recording message in the list holding unit 116 as a recalling target. When the calling device 100 according to the first embodiment is used, it can be automatically and more accurately determined whether the answering machine function has responded or not, so that it can be announced more efficiently even when an emergency response is announced. In addition, the reach to people also improves.

(Third application)
A third application for applying the calling device 100 according to the first embodiment to the initial dunning service of the call center will be described. In the initial reminder business, a call is made to the debtor who is delinquent to repay it. In this application, the list holding unit 116 holds a list of obligor telephone numbers that are delinquent and require initial reminder.

  When the guidance-continuous voice detection mode is used, the message holding unit 118 holds the message content “If you say XX. Are you YY?” For the message type “initial guidance message”. Further, the obligor's surname that can be extracted from the list holding unit 116 is inserted into YY. The message holding unit 118 holds the message content “I will contact you again” for the message type “recorded message”.

  The calling device 100 makes a call to a telephone number of a certain debtor extracted from the list holding unit 116. When the communication device on the called side answers, the calling device 100 sends an initial guidance message “Hello, say XX. Are you YY?” At this time, if the debtor is present at the called party, it is generally silent while the initial guidance message is flowing, and a response such as “Yes” is usually received several tens to several hundreds of milliseconds after the initial guidance message. It is. Therefore, the calling device 100 determines that the response is not a human response when voice is continuously detected within a period of several tens of milliseconds after the initial guidance message starts flowing until the message ends. When the calling device 100 determines that the response is not a human response and further detects “P sound”, the calling device 100 plays a recording message “I will contact you again”. This leaves a message on the answering machine function of the debtor's telephone terminal or the answering machine function of the carrier with which the debtor has contracted.

When using the answering control mode, the message holding unit 118 holds the following question to the called party and the expected response thereto.
Question 1 to the callee: “Hello, say XX. Do you have YY?” (Same as the initial guidance message above)
Expected response 1: “Yes”
Question 2 to the called party: “Do you have YYZZ?” (ZZ is the name of the debtor that can be extracted from the list holding unit 116)
Expected response 2: “Yes”
Question 3 to the called party: “YYZZ, you ’re here.”
Expected response 3: “Yes.
Question 4 to the callee: “I am always indebted to you. This is a WW credit company.”
Expected response 4: “Yes, I am indebted” (There are various answer patterns.)
Question 5 for the callee: “Today is a phone call for guidance. It was a matter of borrowing the other day (month and day), but it was a payment date yesterday, but it was not possible. Could you confirm it? "
Expected response 5: “Oh, yes, I forgot.” (Prepare a grammar in advance according to the expected response pattern.)
Callee's question 6: “I am very sorry for your busy time, but is it possible to receive money today?”
Expected response 6: “Yes. I understand. I will make an immediate payment (remittance)”.
Callee's question 7: “Thank you very much for your busy schedule, thank you. Thank you in the future.”

  The calling device 100 determines that the response is not a human response when the response from the incoming call side deviates from the above-described response flow. When the calling device 100 determines that the response is not a human response and further detects “P sound”, the calling device 100 plays a recording message “I will contact you again”. This leaves a message on the answering machine function of the debtor's telephone terminal or the answering machine function of the carrier with which the debtor has contracted.

(Second Embodiment)
In the first embodiment, a case has been described in which whether or not the response from the communication device on the called side is made by a human and waiting for a “P sound” are performed as separate steps. In the calling device 400 according to the second embodiment, these steps are integrated by analyzing the frequency spectrum in more detail.

  FIG. 9 is a block diagram showing the function and configuration of calling device 400 according to the present embodiment. Each block shown here can be realized by hardware such as a computer (CPU) (central processing unit) and other elements and mechanical devices, and software can be realized by a computer program or the like. Here, The functional block realized by those cooperation is drawn. Therefore, it is understood by those skilled in the art who have touched this specification that these functional blocks can be realized in various forms by a combination of hardware and software.

  The calling device 400 includes a line control unit 102, a calling control unit 104, a recorded message transmission control unit 114, a list holding unit 116, a message holding unit 118, and a spectrum analysis unit 402. The call control unit 104 includes a call target extraction unit 120 and a call instruction unit 122.

  FIG. 10 is a flowchart illustrating an example of a series of processes in the calling device 400. The calling control unit 104 causes the line control unit 102 to automatically make a call to a telephone number included in the telephone number list held in the list holding unit 116 (S230). Here, the calling device 400 may send an initial guidance message. The spectrum analysis unit 402 performs frequency analysis on a signal acquired by the line control unit 102 from the communication device on the called side, and acquires a frequency spectrum (S232). The spectrum analysis unit 402 determines whether or not the acquired frequency spectrum has a peak only at the frequency corresponding to the P sound (S234). When it is determined that only the P sound corresponding frequency has a peak (Y in S234), the spectrum analysis unit 402 determines whether or not such a frequency spectrum continues for the P sound duration (S236). When the P sound duration has continued (Y in S236), the spectrum analysis unit 402 determines that “P sound” has been detected. The recorded message transmission control unit 114 extracts the recorded message from the message holding unit 118 and transmits it to the communication device on the called side (S238). When the transmission of the recording message is completed, the line control unit 102 disconnects the call (S240). When the P sound duration time is not continued (N in S236), the spectrum analysis unit 402 acquires the frequency spectrum again (S232).

If it is not determined in step S234 that there is a peak only in the frequency corresponding to the P sound (N in S234), the spectrum analysis unit 402 determines whether all the components of the audio band are equal to or lower than a predetermined level (S242). ). When it is not determined that all the components of the voice band are equal to or lower than the predetermined level (N in S242), the spectrum analysis unit 402 acquires the frequency spectrum again (S232). When it is determined that all the components of the voice band are equal to or lower than the predetermined level (Y in S242), the spectrum analysis unit 402 determines whether or not such a frequency spectrum continues for a predetermined silence duration (S244). ). When the silent duration is not continued (N in S244), the spectrum analysis unit 402 acquires the frequency spectrum again (S232). When the silent duration is continued (Y in S244), the spectrum analysis unit 402 determines that it is highly possible that the “P sound” has not been detected or that the response from the communication device on the called side is from a human. The silent duration is set longer than the interval between a general answering machine response message and the “P sound”. The spectrum analysis unit 402 performs appropriate automatic response control (S246). Thereafter, the line control unit 102 disconnects the call (S240).
The calling device 400 repeats the above series of processes for other telephone numbers included in the telephone number list held in the list holding unit 116.

  According to the calling device 400 according to the present embodiment, the same operational effects as the operational effects achieved by the calling device 100 according to the first embodiment are exhibited.

  The configuration and operation of the calling device according to the embodiment and the application thereof have been described above. These embodiments are exemplifications, and it is understood by those skilled in the art that various modifications can be made to each component and combination of processes, and such modifications are within the scope of the present invention. .

  The calling device according to the first or second embodiment is not limited to the first, second, and third applications, and can be applied to various other business systems. For example, the calling device according to the first or second embodiment can be applied to a safety confirmation business, a marketing business, or a medical examination date confirmation business.

  2 telephone system, 4 telephone terminal, 6 telephone network, 8 private branch exchange, 10 operator terminal, 12 call center, 100 calling device, 102 line control unit, 104 calling control unit, 106 first discrimination unit, 108 second discrimination unit 110 Transfer control unit 112 Standby unit 114 Recording message transmission control unit 116 List holding unit 118 Message holding unit 400 Calling device

Claims (7)

A line control unit for controlling communication with a communication device via a telephone network;
A call control unit that extracts a telephone number to be called from a list of phone numbers and causes the line control unit to make a call to the telephone number;
A determination unit that determines whether a response from the communication device on the called side is due to a human by analyzing a signal acquired from the communication device on the called side by the line control unit;
When the determination unit determines that the response is not a human response, the line control unit monitors a signal acquired from the called communication device and indicates that recording is started at the called communication device. A standby unit for waiting for a signal corresponding to the sound,
The determination unit causes the line control unit to send a signal corresponding to a question to the called communication device to the called communication device when the line is in a call state in response to the outgoing call, By comparing the response from the called communication device with the expected response to the question, it is determined whether or not the response from the called communication device is human. Call device. A line control unit for controlling communication with a communication device via a telephone network;
A call control unit that extracts a telephone number to be called from a list of phone numbers and causes the line control unit to make a call to the telephone number;
A determination unit that determines whether a response from the communication device on the called side is due to a human by analyzing a signal acquired from the communication device on the called side by the line control unit;
When the determination unit determines that the response is not a human response, the line control unit monitors a signal acquired from the called communication device and indicates that recording is started at the called communication device. A standby unit for waiting for a signal corresponding to the sound,
The determination unit performs emotion analysis on the response from the communication device on the called side, and determines that the response from the communication device on the called side is not a human being when the change in emotion is within a predetermined range. Calling device characterized. A line control unit for controlling communication with a communication device via a telephone network;
A call control unit that extracts a telephone number to be called from a list of phone numbers and causes the line control unit to make a call to the telephone number;
A determination unit that determines whether a response from the communication device on the called side is due to a human by analyzing a signal acquired from the communication device on the called side by the line control unit;
When the determination unit determines that the response is not a human response, the line control unit monitors a signal acquired from the called communication device and indicates that recording is started at the called communication device. A standby unit for waiting for a signal corresponding to the sound,
The waiting unit is configured such that a frequency spectrum having a peak only at a predetermined frequency continues for a predetermined time with respect to a frequency spectrum of a signal acquired by the line control unit from a communication device on the called side. A calling device that detects a signal corresponding to a sound indicating that recording is started in a communication device. When a signal corresponding to a sound indicating that recording is to be started is detected in the call side communication device in the standby unit, a signal corresponding to a predetermined message is sent to the line control unit to the call side communication device. calling device according to claim 1, further comprising a message transmission control unit which sent for 3. A calling method executed by a calling device including a line control unit that controls communication with a communication device via a telephone network,
Extracting a telephone number to be called from a list of telephone numbers and causing the line control unit to make a call to the telephone number;
Determining whether the response from the called communication device is a human response by analyzing the signal acquired from the called communication device by the line control unit;
If it is determined in the determining step that the response is not by a person, the line control unit monitors a signal acquired from the called communication device, and recording is started in the called communication device. Waiting for a signal corresponding to the sound to be indicated ,
In the step of determining, when the line enters a busy state corresponding to the outgoing call, the line control unit sends a signal corresponding to a question to the incoming communication apparatus to the incoming communication apparatus. The response from the called communication device to the question is compared with the expected response to determine whether the response from the called communication device is human. Calling method. A function for controlling communication with a communication device via a telephone network;
A function for extracting a telephone number to be called from a list of telephone numbers and causing the function to be controlled to make a call to the telephone number;
A function for determining whether or not a response from the communication device on the called side is caused by a human by analyzing a signal acquired from the communication device on the called side by the function to be controlled ;
When it is determined in the determining function that the response is not caused by a person, the control function monitors a signal acquired from the called communication device, and recording is started in the called communication device. The computer has a function of waiting for a signal corresponding to the sound to be shown ,
The function for determining is to send a signal corresponding to a question to the communication device on the called side to the communication device on the called side when the line enters a busy state corresponding to the outgoing call. The response from the called communication device to the question is compared with the expected response to determine whether the response from the called communication device is human. Computer program. A function for controlling communication with a communication device via a telephone network;
A function for extracting a telephone number to be called from a list of telephone numbers and causing the function to be controlled to make a call to the telephone number;
A function for determining whether or not a response from the communication device on the called side is caused by a human by analyzing a signal acquired from the communication device on the called side by the function to be controlled;
When it is determined in the determining function that the response is not caused by a person, the control function monitors a signal acquired from the called communication device, and recording is started in the called communication device. The computer has a function of waiting for a signal corresponding to the sound to be shown,
The waiting function is determined on the condition that a frequency spectrum having a peak only at a predetermined frequency continues for a predetermined time with respect to a frequency spectrum of a signal acquired by the function to be controlled from a communication device on the called side. A computer program for detecting a signal corresponding to a sound indicating that recording is started in a communication device.