JP2019020589A - Voice recognition system and processing stop method in the same - Google Patents

Abstract

To provide a voice recognition system capable of cancelling a stop instruction even if the stop instruction of processing in execution is erroneously recognized.SOLUTION: A voice recognition system 10 comprises: a voice recognition section 11 for constantly analyzing peripheral voice and recognizing a voice instruction; and a processing execution section 13 for performing processing corresponding to the voice instruction. The processing execution section 13 comprises: a stop reservation setting section 131 for reserving stop processing in execution when the voice instruction is the stop instruction of the processing in execution; a stop reservation execution section 132 for performing the reserved stop processing when a previously determined time elapses from the reservation of the stop processing; a user intention estimation section 134 for estimating intention of a user from information except for the voice instruction; and a stop reservation cancel section 133 for cancelling the reservation of the stop processing when the user is estimated to have no stop intention of the processing in execution before the previously determined time elapses from the reservation of the stop processing.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a speech recognition system that constantly analyzes speech.

  2. Description of the Related Art A voice recognition system that recognizes a user's voice instruction by constantly analyzing voice emitted in the surroundings is known. In such a voice recognition system, there is a problem of preventing erroneous recognition that a user's utterance is unintentionally recognized as a voice instruction.

  For example, in Patent Document 1 below, it is determined whether or not the user has an intention to perform a voice instruction (voice operation), and when it is determined that the intention is not present, it is compared with a case where it is determined that there is an intention. Thus, there has been proposed a speech recognition system that changes the control mode of the controlled device in a direction that does not make the user aware of it (a direction that does not disturb the user). For example, if the voice recognition system fails in voice recognition, the voice recognition system outputs a voice message “Please speak again” if the user judges that the voice is intended to be voiced. If there is no intention, do nothing.

International Publication No. 2016/051519

  In the speech recognition system of Patent Document 1, the occurrence of misrecognition is suppressed by determining the user's intention, but the determination is made when the user utters, and the behavior of the user after the utterance is not considered. . For this reason, once a voice instruction is erroneously recognized, it cannot be canceled. In particular, if a voice instruction for stopping (including ending or interrupting) a process being executed is erroneously recognized and the process is stopped against the user's intention, a problem such as loss of unsaved data occurs. For example, a case may be considered in which a voice recognition system mounted on a vehicle misidentifies a passenger's utterance “stop the car” as a voice instruction to stop the process being executed.

  The present invention has been made to solve the above-described problems, and provides a voice recognition system capable of canceling a stop instruction even if the stop instruction for the process being executed is erroneously recognized. With the goal.

  The voice recognition system according to the present invention includes a voice recognition unit that recognizes a voice instruction by constantly analyzing surrounding voices, and a process execution unit that executes a process according to the voice instruction recognized by the voice recognition unit. The process execution unit includes a stop reservation setting unit for reserving a stop process for a process being executed when a voice instruction is an instruction to stop the process being executed, and a predetermined time has elapsed from the reservation of the stop process. Then, the stop reservation execution unit that executes the reserved stop process, the user intention estimation unit that estimates the user's intention from information other than the voice instruction, and before a predetermined time elapses from the reservation of the stop process, When the user intention estimation unit estimates that the user does not intend to stop the process being executed, the user intention estimation unit includes a stop reservation cancellation unit that cancels the reservation of the stop process.

  The speech recognition system according to the present invention reserves a stop process so that it stops after a predetermined time, instead of immediately stopping the process being executed, when it recognizes a stop instruction of the process being executed. . However, if it is estimated that the user does not intend to stop the process being executed before the time has elapsed, the speech recognition system cancels the reservation for the stop process. Therefore, even when the stop instruction is recognized by mistake, it can be canceled afterwards according to the user's intention.

1 is a functional block diagram illustrating a configuration of a speech recognition system according to Embodiment 1. FIG. 4 is a flowchart illustrating an operation for a voice instruction of the voice recognition system according to the first embodiment. 4 is a flowchart illustrating an operation for manual operation of the voice recognition system according to the first embodiment. 4 is a flowchart illustrating an operation related to a reserved stop process of the voice recognition system according to the first embodiment. It is a figure which shows the example of the hardware constitutions of a speech recognition system. It is a figure which shows the example of the hardware constitutions of a speech recognition system. 6 is a flowchart showing a stop process of an ongoing process in the speech recognition system according to the second embodiment. 14 is a flowchart illustrating an operation for a voice instruction of the voice recognition system according to the third embodiment. It is a functional block diagram which shows the structure of the speech recognition system which concerns on Embodiment 4. FIG. 10 is a diagram for explaining a modification of the voice recognition system according to the fourth embodiment.

<Embodiment 1>
FIG. 1 is a functional block diagram showing the configuration of the speech recognition system 10 according to the first embodiment. As shown in FIG. 1, the voice recognition system 10 includes a microphone 1 for a user to input voice into the voice recognition system 10, a manual operation input device 2 for a user to operate the voice recognition system 10, An information output device 3 is connected to the voice recognition system 10 for presenting information to the user.

  The manual operation input device 2 includes, for example, a keyboard, a touch pad, and a push button. The information output device 3 includes, for example, a screen that displays images and text messages, a speaker that outputs voice messages and sound effects, and the like. Alternatively, the manual operation input device 2 and the information output device 3 may be configured as one touch panel monitor by arranging a touch pad as the manual operation input device 2 on the screen of the information output device 3.

  The voice recognition system 10 includes a voice recognition unit 11, a manual operation recognition unit 12, and a process execution unit 13. The voice recognition unit 11 recognizes a user's voice instruction by constantly analyzing the surrounding voice acquired by the microphone 1. The manual operation recognition unit 12 recognizes an operation input to the manual operation input device 2 (hereinafter referred to as “manual operation”). The process execution unit 13 executes a process corresponding to the voice instruction recognized by the voice recognition unit 11 or the manual operation recognized by the manual operation recognition unit 12.

  As illustrated in FIG. 1, the process execution unit 13 includes a stop reservation setting unit 131, a stop reservation execution unit 132, a stop reservation cancellation unit 133, and a user intention estimation unit 134.

  When the voice instruction acquired from the voice recognition unit 11 is an instruction to stop the process being executed, the stop reservation setting unit 131 is configured so that the process related to the stop is executed after a predetermined time. Schedule stop processing. In the following, the process being executed is “in-process”, the voice instruction to stop the process being executed is “stop instruction”, the process to stop the process in progress is “stop process”, and the process from reservation to execution The time (the above “predetermined time”) is referred to as “stop hold time”. In this specification, “stop” of the process means “end” to stop the process completely (stop that cannot be resumed from the stop point) or “suspend” to temporarily stop (restart from the stop point) Including a wide range of concepts such as stoppages).

  The stop reservation execution unit 132 measures the elapsed time from the time when the stop reservation setting unit 131 sets the stop processing reservation, and performs the reserved stop processing when the stop hold time has elapsed from the stop processing reservation. Run.

  The user intention estimation unit 134 estimates from the information other than the voice instruction recognized by the voice recognition unit 11 whether or not the user's intention, specifically, the user has an intention to stop the running process. In the present embodiment, when the user performs some manual operation using the manual operation input device 2, the user intention estimation unit 134 estimates that the user has no intention to stop the running process. On the other hand, if no manual operation has been performed since the stop process of the ongoing process is reserved, the user intention estimation unit 134 estimates that the user has an intention to stop the ongoing process.

  If the user intention estimation unit 134 estimates that the user has no intention to stop the running process before the stop hold time elapses from the reservation of the stop process, the stop reservation canceling unit 133 cancels the reservation of the stop process.

  Here, the operation of the speech recognition system 10 according to Embodiment 1 will be described using the flowcharts of FIGS.

  First, the operation of the voice recognition system 10 in response to voice instructions will be described with reference to FIG. When the voice recognition system 10 is activated, the voice recognition unit 11 analyzes the voice acquired by the microphone 1 (step S101). If the voice recognition unit 11 does not recognize the user's voice instruction (NO in step S102), step S101 is repeated. That is, the surrounding voice analysis by the voice recognition unit 11 is always performed.

  When the voice recognition unit 11 recognizes the user's voice instruction (YES in step S102), the recognition result is transmitted to the process execution unit 13, and the process execution unit 13 indicates that the voice instruction is an instruction to stop the ongoing process. It is confirmed whether or not there is (step S103).

  When the voice instruction of the user recognized by the voice recognition unit 11 is other than a stop instruction for the process being executed (NO in step S103), the process execution unit 13 performs a process according to the voice instruction. Execute (step S104) and return to step S101. Depending on the content of the process executed by the process execution unit 13, an image or sound is output from the information output device 3.

  On the other hand, when the voice instruction of the user recognized by the voice recognition unit 11 is a stop instruction for the ongoing process (NO in step S103), after the stop process of the ongoing process has elapsed for the stop hold time. The stop reservation setting unit 131 reserves a stop process so as to be executed (step S105), and the process returns to step S101.

  As described above, when the voice recognition system 10 recognizes the voice instruction of the user, the voice recognition system 10 immediately executes a process corresponding to the voice instruction. However, when the voice instruction is an instruction to stop the ongoing process, it is being executed. Instead of stopping the process immediately, stop the process after the stop hold time has elapsed.

  Next, the operation | movement with respect to the manual operation of the speech recognition system 10 is demonstrated, referring FIG. When the voice recognition system 10 is activated, the manual operation recognition unit 12 confirms whether or not a user's manual operation is input to the manual operation input device 2 (step S201). When a manual operation is input to the manual operation input device 2 (YES in step S201), the manual operation recognition unit 12 recognizes the content of the manual operation, and the recognition result is transmitted to the process execution unit 13. The process execution part 13 performs the process according to the said manual operation (step S202), and returns to step S201. Note that step S201 is repeated while no manual operation is input to the manual operation input device 2 (NO in step S201).

  As described above, the voice recognition system 10 immediately executes a process (including an ongoing process including a stop process) according to a manual operation input by the user.

  The flow in FIG. 2 and the flow in FIG. 3 are performed in parallel. Further, there is no substantial difference between the process executed by the process execution unit 13 in step S104 of FIG. 2 and the process executed by the process execution unit 13 in step S202 of FIG. For example, the process executed by the process execution unit 13 in response to a voice instruction can be stopped by a manual operation. Conversely, the process executed by the process execution unit 13 in response to a manual operation is stopped by a voice operation. (In this case, the stop process is reserved).

  Next, with reference to FIG. 4, an operation related to the reserved stop process of the speech recognition system 10 will be described. The flow in FIG. 4 is executed by the stop reservation executing unit 132, the stop reservation canceling unit 133, and the user intention estimating unit 134 of the process executing unit 13.

  When the speech recognition system 10 is activated, the stop reservation execution unit 132 confirms whether there is a reservation for a valid stop process set by the stop reservation setting unit 131 (step S301). A reservation that has already been canceled, a reservation that has already been executed, or a reservation for a process that has already stopped is not valid. If there is no valid stop process reservation (NO in step S301), step S301 is repeated until a new reservation is set.

  If there is a reservation for a valid stop process (YES in step S301), whether the user intention estimation unit 134 has the intention to stop the ongoing process based on the manual operation recognition result by the manual operation recognition unit 12 Whether or not is estimated (step S302). That is, the user intention estimation unit 134 estimates that the user has no intention to stop the ongoing process if any manual operation is performed by the user. On the other hand, if no manual operation has been performed since the stop process of the ongoing process is reserved, the user intention estimation unit 134 estimates that the user has an intention to stop the ongoing process.

  When the user intention estimation unit 134 estimates that the user does not intend to stop the process being executed (NO in step S303), the stop reservation cancellation unit 133 cancels the reservation of the stop process (step S304), Return to step S301.

  On the other hand, if the user intention estimation unit 134 estimates that the user has an intention to stop the process being executed (YES in step S303), the stop reservation execution unit 132 has elapsed the stop hold time from the reservation of the stop process. It is confirmed whether or not it has been done (step S305). At this time, if the stop hold time has elapsed (YES in step S305), the stop reservation execution unit 132 executes the reserved stop process (step S306), and the process returns to step S301. However, if the stop hold time has not elapsed (NO in step S305), the process returns to step S301 while the stop process reservation is maintained.

  As described above, the speech recognition system 10 according to Embodiment 1 does not stop the in-execution process immediately but stops after the stop hold time has elapsed, even if an instruction to stop the in-execution process is recognized. Let However, if it is presumed that the user does not intend to stop the ongoing process before the stop suspension time elapses, the reservation of the stop process is canceled. Therefore, even when the stop instruction is recognized by mistake, it can be canceled afterwards according to the user's intention.

  The application of the voice recognition system 10 according to the present invention is not particularly limited, and can be widely applied to electronic devices that can input voice instructions, such as a PC (personal computer), a mobile phone, a smartphone, and a navigation device.

  FIG. 5 and FIG. 6 are diagrams each illustrating an example of a hardware configuration of the speech recognition system 10. Each element (speech recognition unit 11, manual operation recognition unit 12, and process execution unit 13) of the speech recognition system 10 illustrated in FIG. 1 is realized by, for example, a processing circuit 50 illustrated in FIG. That is, the processing circuit 50 recognizes a voice instruction by constantly analyzing surrounding sounds, a manual operation recognition unit 12 that recognizes a user's manual operation, and a voice recognized by the voice recognition unit 11. And a process execution unit 13 that executes a process according to the instruction. The process execution unit 13 also has a stop reservation setting unit 131 for reserving a stop process for the ongoing process when the voice instruction is a stop instruction for the ongoing process, and a stop hold time from the stop process reservation. When the time has elapsed, the stop reservation execution unit 132 that executes the reserved stop process, the user intention estimation unit 134 that estimates the user's intention from information other than the voice instruction, and before the stop hold time elapses from the reservation of the stop process When the user intention estimation unit 134 estimates that the user does not intend to stop the process being executed, a stop reservation cancellation unit 133 that cancels the reservation of the stop process is provided. Dedicated hardware may be applied to the processing circuit 50, or a processor (a central processing unit (CPU), a processing device, an arithmetic device, a microprocessor, a micro processor) that executes a program stored in a memory A computer or a DSP (Digital Signal Processor) may also be applied.

  When the processing circuit 50 is dedicated hardware, the processing circuit 50 includes, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), and an FPGA (Field-Programmable). Gate Array) or a combination of these. Each of the functions of each element of the speech recognition system 10 may be realized by a plurality of processing circuits, or these functions may be realized by a single processing circuit.

  FIG. 6 shows a hardware configuration of the speech recognition system 10 when the processing circuit 50 is configured using a processor. In this case, the function of each element of the speech recognition system 10 is realized by a combination of software and the like (software, firmware, or software and firmware). Software or the like is described as a program and stored in the memory 52. The processor 51 as the processing circuit 50 implements the functions of the respective units by reading out and executing the program stored in the memory 52. That is, the voice recognition system 10 performs processing that recognizes the voice instruction of the user by constantly analyzing surrounding voices and processing that corresponds to the recognized voice instructions when executed by the processing circuit 50. Pre-determined from a process for reserving a stop process for an ongoing process when the voice instruction is a stop instruction for a process being executed, a process for estimating a user's intention from information other than a voice instruction, and a reservation for a stop process A process for executing a reserved stop process when a predetermined time elapses, and a stop process when it is estimated that the user does not intend to stop the process being executed before a predetermined time elapses from the reservation of the stop process. And a memory 52 for storing a program to be executed as a result. In other words, it can be said that this program causes a computer to execute the operation procedure and method of each element of the speech recognition system 10.

  Here, the memory 52 is nonvolatile or non-volatile such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), etc. Volatile semiconductor memory, HDD (Hard Disk Drive), magnetic disk, flexible disk, optical disk, compact disk, mini disk, DVD (Digital Versatile Disc) and its drive device, etc., or any storage media used in the future May be.

  The configuration in which the function of each element of the voice recognition system 10 is realized by either hardware or software has been described above. However, the present invention is not limited to this, and a configuration may be adopted in which some elements of the speech recognition system 10 are realized by dedicated hardware and another part is realized by software or the like. For example, the functions of some elements are realized by the processing circuit 50 as dedicated hardware, and the processing circuit 50 as the processor 51 reads a program stored in the memory 52 for the other some elements. The function can be realized by executing.

  As described above, the speech recognition system 10 can realize the above-described functions by hardware, software, or the like, or a combination thereof.

<Embodiment 2>
It is assumed that some of the processes executed by the process execution unit 13 ask the user for permission to stop before stopping. The voice recognition system 10 according to the second embodiment requests permission from the user when stopping the ongoing process according to the user's manual operation, but omits the request when stopping the process according to the voice instruction.

  FIG. 7 is a flowchart showing a stop process of the ongoing process in the speech recognition system 10 according to the second embodiment. That is, the flow of FIG. 7 is performed when the stop process of the running process is performed according to the manual operation in step S202 of FIG. 3, or when the reserved stop process is executed in step S306 of FIG. To be done.

  When the process execution unit 13 starts the stop process of the process being executed, the process execution unit 13 checks whether the process being executed has a flow for requesting the user to stop (a stop permission request flow). (Step S401). If the in-execution process does not have a stop permission request flow (NO in step S401), the process execution unit 13 stops the in-execution process as it is (step S405) and ends the flow of FIG.

  On the other hand, if the ongoing process to be stopped has a stop permission request flow (YES in step S401), the process execution unit 13 determines whether the current stop process is being performed according to the user's manual operation. Alternatively, it is confirmed whether or not it is performed according to an instruction from the stop reservation execution unit 132 (step S402). In other words, in step S402, it is confirmed whether the current stop process is performed in step S202 in FIG. 3 or in step S306 in FIG.

  When the current stop process is performed according to the user's manual operation (NO in step S402), the process execution unit 13 executes a stop permission request flow for the process being executed (step S403). In the stop permission request flow, the process execution unit 13 uses the information output device 3 to ask the user for permission to stop the process being executed. For example, a voice message such as “Do you want to stop the XX process?” Is output from the speaker of the information output device 3, or a similar character message is displayed on the screen of the information output device 3. The user can make a response as to whether or not to stop the ongoing process by voice instruction or manual operation.

  If the process execution unit 13 obtains the user's permission in the stop permission request flow (YES in step S404), the process execution unit 13 stops the execution process to be stopped (step S405) and ends the flow of FIG. However, if the user's permission cannot be obtained (NO in step S404), the process in FIG. 7 is terminated without stopping the running process.

  On the other hand, when the current stop process is performed according to an instruction from the stop reservation execution unit 132 (YES in step S402), the process execution unit 13 omits the execution of the stop permission request flow (step S406), The in-execution process to be stopped is stopped (step S405), and the flow of FIG. 7 is ended.

  As described above, in the present embodiment, the process execution unit 13 requests the user's permission when performing the stop process of the in-execution process in response to a manual operation, and performs the in-execution process according to an instruction from the stop reservation execution unit 132. When the stop process is performed, the user permission is not requested. This eliminates the need for the user to perform an additional voice instruction or manual operation between the time when the in-execution process is instructed and the time when the in-execution process is stopped, thereby improving the convenience of the voice recognition system 10. To do.

<Embodiment 3>
In the speech recognition system 10 according to the first and second embodiments, the user can cancel the reservation until the stop hold time elapses after the stop process of the process being executed is reserved. However, if the user does not notice that the stop process has been reserved, it is also conceivable that the executing process stops against the user's intention. Therefore, in the third embodiment, when the speech recognition system 10 reserves the stop process of the ongoing process, it notifies the user.

  FIG. 8 is a diagram illustrating an operation in response to a voice instruction of the voice recognition system 10 according to the third embodiment. In the flow of FIG. 8, step S106 for notifying the user that the stop process is reserved is added to the flow of FIG. 2 after step S105 for reserving the stop process.

  In step S <b> 106, the stop reservation setting unit 131 that has set reservation for stop processing notifies the user that the stop processing has been reserved using the information output device 3. For example, the stop reservation setting unit 131 outputs a voice message from the speaker of the information output device 3 such as “the stop of the XX process is reserved” or “the data is automatically saved after XX seconds and ends”. Or a similar character message is displayed on the screen of the information output device 3. When the stop process is reserved, the screen color of the information output device 3 may be changed, the icon color may be changed, or the icon may be blinked. In addition, an indicator indicating the remaining time until the reserved stop process is executed may be displayed on the screen of the information output device 3.

  According to the present embodiment, since the user can be aware that the stop process is reserved, it is possible to prevent the ongoing process from stopping against the user's intention.

<Embodiment 4>
FIG. 9 is a functional block diagram showing the configuration of the speech recognition system 10 according to the fourth embodiment. The voice recognition system 10 in FIG. 9 is configured to include a user behavior information acquisition unit 14 instead of the manual operation recognition unit 12 in the configuration of FIG. The user behavior information acquisition unit 14 is connected to a camera 4 that captures the user's video and a sensor 5 that acquires the user's biological information. Only one of the camera 4 and the sensor 5 may be used.

  The user behavior information acquisition unit 14 acquires user behavior information that is information indicating a user's behavior from the camera 4 or the sensor 5. The contents of the user behavior information include, for example, the user's position, action (gesture), face direction, mouth movement, line-of-sight direction, and breathing action during speech.

  Further, the user intention estimation unit 134 according to the present embodiment estimates whether or not the user has an intention to stop the running process from the user behavior information acquired by the user behavior information acquisition unit 14. For example, when the direction of the user's line of sight or the direction of the face is facing the screen of the information output device 3 or when the user performs an action of reaching for the manual operation input device 2, the user (for voice instruction) B) When the user opens his / her mouth or inhales a lot, the user intention estimation unit 134 estimates that the user has no intention to stop the ongoing process. Even if the user behaves as described above, if the user's position is more than a certain distance from the microphone 1, the manual operation input device 2, or the information output device 3, When the user 2 or the information output device 3 leaves, for example, the user intention estimation unit 134 may estimate that the user has an intention to stop the ongoing process.

  The operation of the speech recognition system 10 according to the fourth embodiment is basically the same as that of the first embodiment (FIGS. 2 to 4). However, in step S302 in FIG. 4, the user intention estimation unit 134 estimates whether or not the user has an intention to stop the ongoing process from the user behavior information.

  In the fourth embodiment, although the criterion for determining whether or not the user has the intention to stop the running process is different from that in the first embodiment, the same effect as in the first embodiment can be obtained.

  The fourth embodiment can be combined with the first to third embodiments. That is, as shown in FIG. 10, the voice recognition system 10 may be provided with both the manual operation recognition unit 12 and the user behavior information acquisition unit 14. In that case, the user intention estimation unit 134 performs processing that is being executed by the user both when the user's manual operation is input to the manual operation input device 2 and when the user's behavior satisfies a predetermined condition. It is estimated that there is no intention to stop.

<Modification>
In the present invention, the voice that the voice recognition unit 11 determines to be an instruction to stop the ongoing process is basically a language such as “stop”, “stop”, “end”, etc. Non-language speech.

  For example, if a user hears a child's cry while using an Internet mail order service using a PC, the user will leave the PC and take care of the child. At this time, when the user forgets to temporarily save the shopping data and the system side of the mail order service times out, the information in the middle of shopping is lost, and when shopping is resumed, it is necessary to redo the shopping from the beginning.

  In the example of the Internet mail order service described above, if the voice recognition system 10 that determines that a child's cry is an instruction to stop is installed in the PC, the shopping function process is stopped when the voice recognition system 10 detects a child's cry. Is reserved. Thereafter, when the stop hold time elapses, a stop process (including a process for temporarily storing shopping data) is automatically executed on the PC side, so that it is possible to avoid the disappearance of information during shopping. If the user finishes taking care of the child and performs a manual operation or a specific gesture before the stop hold time elapses, the reservation of the stop processing for the shopping function processing is canceled and the shopping can be continued.

  As non-verbal speech that is determined to be an instruction to stop processing during execution, in addition to a child's cry, a pet's cry, an earthquake impact sound, and the like can be considered.

  It should be noted that the present invention can be freely combined with each other within the scope of the invention, and each embodiment can be appropriately modified or omitted.

  1 microphone, 2 hand operation input device, 3 information output device, 4 camera, 5 sensor, 10 speech recognition system, 11 speech recognition unit, 12 hand operation recognition unit, 13 process execution unit, 14 user behavior information acquisition unit, 131 stop Reservation setting part, 132 Stop reservation execution part, 133 Stop reservation cancellation part, 134 User intention estimation part, 50 Processing circuit, 51 Processor, 52 Memory.

Claims (7)

A voice recognition unit that recognizes voice instructions by constantly analyzing surrounding voices;
A process execution unit that executes a process according to the voice instruction recognized by the voice recognition unit;
The process execution unit
A stop reservation setting unit for reserving a stop process of the running process when the voice instruction is a stop instruction of the running process;
When a predetermined time has elapsed from the reservation of the stop process, a stop reservation execution unit that executes the reserved stop process;
A user intention estimation unit that estimates a user's intention from information other than the voice instruction;
If the user intention estimation unit estimates that the user does not intend to stop the process being executed before the predetermined time has elapsed since the reservation of the stop process, the reservation of the stop process is canceled. A stop reservation canceling section,
A speech recognition system comprising: A manual operation recognition unit that recognizes the user's manual operation;
The user intention estimation unit estimates that the user does not intend to stop the process being executed when the manual operation is recognized before the predetermined time elapses from the reservation of the stop process.
The speech recognition system according to claim 1. A manual operation recognition unit that recognizes the user's manual operation;
The process execution unit requests permission of the user when performing the process stop process in response to the manual operation, and performs the process stop process according to an instruction from the stop reservation execution unit. The voice recognition system according to claim 1 or 2, wherein when performing the operation, permission of the user is not requested. A user behavior information acquisition unit for acquiring information indicating the user behavior;
If the user behavior satisfies a predetermined condition before the predetermined time has elapsed since the reservation of the stop process, the user intention estimation unit may stop the user from stopping the process being executed. Presumed not,
The voice recognition system according to any one of claims 1 to 3. The speech recognition system according to any one of claims 1 to 4, wherein the stop reservation setting unit notifies the user when the stop process is reserved. The voice recognition system according to claim 5, wherein the notification to the user is a display or voice notification. A method for stopping processing in a voice recognition system in a voice recognition system,
The voice recognition unit of the voice recognition system recognizes voice instructions by constantly analyzing surrounding voices,
A processing execution unit of the voice recognition system executes a process according to the voice instruction recognized by the voice recognition unit;
When the voice instruction is an instruction to stop the process being executed, the stop reservation setting unit of the voice recognition system reserves the stop process of the process being executed,
The user intention estimation unit of the voice recognition system estimates the user's intention from information other than the voice instruction,
When a predetermined time has elapsed since the reservation of the stop process, the stop reservation execution unit of the voice recognition system executes the reserved stop process,
If the user intention estimation unit estimates that the user does not intend to stop the process being executed before the predetermined time has elapsed since the reservation of the stop process, the stop reservation of the speech recognition system is performed. The cancellation unit cancels the reservation for the stop process,
How to stop processing.

Images