JP2020060730A - Voice recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice recognition device with improved operability of device control by voice recognition.
A voice recognition device controls a device according to an input voice. The voice recognition device accepts a voice input, performs voice recognition, acquires a state of the voice recognition unit that identifies the utterance content and the device, and identifies a scene determined based on the state as a current scene from a plurality of scenes. The scene identification unit, a storage unit that stores in advance one or more combinations of utterance content and control content associated with each of a plurality of scenes, and the storage unit is referred to, and the voice recognition unit identifies the current scene. If the utterance content is associated, the device control unit that executes the control content combined with the utterance content identified by the voice recognition unit associated with the current scene is provided.
[Selection diagram] Figure 2

Description

  The present invention relates to a voice recognition device mounted on a vehicle or the like.

  Various voice recognition techniques have been proposed (Patent Documents 1 and 2). In particular, Patent Document 1 discloses a voice wakeup technique for activating various equipment by uttering a specific wakeup word.

JP, 2017-129860, A JP, 2018-87838, A

  When the voice wake-up technology is applied to an in-vehicle device, it is convenient for the user to operate the device while driving because the user does not have to take his / her hand off the steering wheel. For example, by applying voice wake-up technology to an in-vehicle device that communicates with a user's mobile phone in the vehicle using short-range wireless communication, and relays call voice, etc. Can be responded to.

  In this case, it is conceivable that the subsequent voice input can be accepted by the utterance of the specific wakeup word, and then the incoming call response instruction is given by the specific utterance corresponding to the incoming call response. However, in this case, it is necessary to utter twice, and it is necessary to wait for activation of the voice recognition processing between utterances. As described above, there is room for improvement in controlling the in-vehicle device or the like by voice recognition from the viewpoint of improving operability.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a voice recognition device with improved operability of device control by voice recognition.

  In order to solve the above-mentioned problems, one aspect of the present invention is a voice recognition device capable of controlling a device in accordance with an input voice, which receives voice input, performs voice recognition, and outputs speech content. A voice recognition unit that identifies a device, a scene identification unit that obtains the state of a device, and determines a scene determined based on the state as a current scene, and a scene identification unit that identifies from a plurality of scenes, and one or more utterance contents for each of the plurality of scenes. If the utterance content specified by the voice recognition unit is associated with the current scene by referring to the storage unit and the storage unit that stores the combination of the control contents in advance, it is associated with the current scene. And a device control unit that executes control content combined with the utterance content specified by the voice recognition unit.

  INDUSTRIAL APPLICABILITY The present invention can provide a voice recognition device with improved operability of device control by voice recognition.

The figure which shows the functional block of the speech recognition apparatus which concerns on one Embodiment of this invention. The flowchart which shows the example of a process of the speech recognition apparatus which concerns on one Embodiment of this invention.

(Embodiment)
A voice recognition device according to an embodiment of the present invention not only starts voice recognition processing by one utterance, but also specifies the control content of a control target device according to a scene representing an operation state of the control target device. Then run. This eliminates the need for the two utterances and the waiting time between them, thus improving the operability. Hereinafter, the present embodiment will be described in detail with reference to the drawings.

<Structure>
FIG. 1 shows a functional block diagram of the voice recognition device 10 according to the present embodiment. The voice recognition device 10 is mounted on a vehicle and includes a voice recognition unit 11, a scene identification unit 12, a storage unit 13, and a device control unit 14. As an example, the voice recognition device 10 is connected to the microphone 20, the first device 31, and the second device 32 mounted on the vehicle by wire, and is close to the third device 33 existing in the vehicle compartment such as BlueTooth (registered trademark). Communication is possible by distance wireless communication. The microphone 20 picks up the user's voice and inputs it to the voice recognition device 10. The first device 31, the second device 32, and the third device 33 are devices to be controlled by the voice recognition device 10, and are, for example, an audio device, an air conditioner, and a mobile phone. The type and number of control target devices are not limited to this.

  The voice recognition unit 11 accepts a voice input from the microphone 20, extracts a phoneme string from the input voice, and specifies the utterance content.

  The scene identification unit 12 acquires predetermined information indicating the operating state of the control target device such as the first device 31, the second device 32, and the third device 33, for example, and identifies the current scene based on this information. To do. If necessary for identifying the scene, information such as the power supply state and speed of the vehicle may be acquired as predetermined information from a device or sensor other than the control target device. The scene is, for example, information indicating the operating status of the control target device, and has a plurality of predetermined types.

  The storage unit 13 stores the utterance content in association with each of the plurality of scenes in advance.

  The device control unit 14 refers to the storage unit 13 and acquires the utterance content associated with the scene identified by the scene identification unit 12. Further, the speech content specified by the voice recognition unit 11 is acquired. These utterance contents are compared, and if they match, the control target device is controlled according to the identified scene.

<Process>
FIG. 2 is a flowchart showing an example of processing of the voice recognition device 10. An example of the process will be described with reference to FIG. The present flowchart is started by voice input from the microphone 20.

  (Step S101): The voice recognition unit 11 extracts a phoneme string from the input voice and specifies the utterance content.

  (Step S102): The scene identification unit 12 obtains information such as an operation state from a control target device such as the first device 31, the second device 32, and the third device 33, and identifies the scene.

  (Step S103): The device control section 14 refers to the storage section 13 and compares the utterance content associated with the scene identified in step S102 with the utterance content identified in step S101. If any of the utterance contents associated with the scene identified in step S102 and the utterance contents identified in step S101 match, the process proceeds to step S104, and if they do not match, the process of this flowchart ends. That is, if the utterance content identified in step S101 is associated with the current scene, the process proceeds to step S104, and if not, the process of this flowchart ends.

  Here, an example of scenes and utterance contents stored in the storage unit 13 will be described. Table 1 shows an example of the stored contents of the storage unit 13. In this example, the first device 31, the second device 32, and the third device 33 are an audio device, an air conditioner, and a mobile phone, respectively. Further, in this example, the storage unit 13 further stores one or more combinations of utterance content and control content for each scene.

  In step S102, the scene identification unit 12 acquires, for example, the operating states of the audio device, the air conditioner, the mobile phone, etc., and identifies any of the scenes in Table 1 as applicable. If a plurality of scenes are applicable, one may be selected based on a predetermined priority, for example.

  For example, when the scene specified in step S102 is "incoming mobile phone call" and the utterance content specified in step S101 is "hello", "what if" is associated with the scene "incoming mobile phone call". Since it is the utterance content, as a result of the processing in step S103, the process proceeds to step S104.

  (Step S104): The device control unit 14 performs control based on the utterance content identified in step S101 according to the scene identified in step S102. That is, the control associated with the utterance content identified in step S101 is executed by referring to the combination associated with the current scene. The device control unit 14 can identify the control content by referring to the storage unit 13. For example, when the scene specified in step S102 is “incoming mobile phone call” and the utterance content specified in step S101 is “hello”, the device control unit 14 causes the third device 33, which is a mobile phone, to access the third device 33. On the other hand, as shown in Table 1, the instruction to answer the incoming call is issued. Then, the process of this flowchart is completed.

  The example of the process of the voice recognition device 10 has been described above, but this process is repeatedly executed each time a voice is input from the microphone 20.

  If two or more utterance contents and control contents are associated with one scene, a plurality of types of control can be performed in that scene. However, the present invention is not limited to this, and one utterance content and one control content may be associated with one scene. Further, in a certain scene, one utterance content may be associated with two or more control contents, or two or more utterance contents may be associated with the same control content.

  Further, the voice recognition device 10 may be provided with a control unit that presents the contents of the above-mentioned control and the result thereof to the user by using a GUI or a speaker like a touch display device in the vehicle. Also, for example, after responding to an incoming call of the mobile phone, the control unit relays the call voice between the mobile phone and the microphone 20 and the speaker while the call continues, and during that time, it is based on the utterance content as described above. Device control may be suppressed.

<Effect>
According to the present invention, it is possible to instruct the control content of the control target device by the utterance of the utterance content determined according to the specified scene. That is, in the past, first, the first utterance was set to the state in which the subsequent voice input can be accepted, and the second utterance was used to instruct the device control content. Device control can be performed. Further, by setting the utterance content according to the scene, it is possible to instruct the desired control content with a relatively short intuitive utterance. As a result, the operability of device control by the voice recognition device can be improved.

(Modification)
In the above example, the scene identifying unit 12 selects one of the plurality of scenes in step S102 when the plurality of scenes correspond, but selects two or more of the plurality of corresponding scenes. May be. In this case, as shown in Table 2 below, utterance contents that do not overlap between scenes are further provided, and device control based on utterance contents that overlap among the selected scenes may be invalidated, and device control based on utterance contents that do not overlap may be enabled. . For example, when both the audio device and the air conditioner are operating and the scene specifying unit 12 selects the scenes “audio device is operating” and “air conditioner is operating”, the device control unit 14 causes the scenes to be selected. Control based on the utterances of "Agete" and "Sagete" that are duplicated between each other shall not be performed. Control based on the utterance "Sagete" should be performed. As a result, erroneous control can be suppressed even when a plurality of devices such as an audio device and an air conditioner are controlled at the same time.

  As described above, the utterance content and the processing content by the scene identification unit 12 and the device control unit 14 can be appropriately designed according to the type of the control target device, the control content, and the like. Further, the types of scenes may be provided for one control target device according to a plurality of operation states thereof, or may be provided for a combination of operation states of two or more control target devices.

(Reference example)
In each of the above examples, the utterance content for device control is associated with each scene, but a plurality of scenes may not be provided as shown in Table 3 below, and only the utterance content and the control content may be associated with each other. . In this case, the above-mentioned scene specifying unit 12 is not provided, the voice recognition unit 11 receives voice input and performs voice recognition to specify the utterance content, and the storage unit 13 stores two or more utterances as shown in Table 3. The contents and the control contents are stored in association with each other, and when the device control unit 14 matches the utterance contents stored in the storage unit 13 with the utterance contents specified by the voice recognition unit, they are associated with the utterance contents. It suffices to execute the controlled contents. Note that the device control unit 14 acquires the operating state of the control target device when performing control assuming that the control target device is in the predetermined state, and does not execute the control when the control target device is not in the predetermined state. Alternatively, if the device to be controlled is not in a predetermined state, the control instruction based on that state may not be accepted.

  Although the embodiments of the present invention have been described above, the present invention is directed to a voice recognition device, a voice recognition method executed by a computer of the voice recognition device, a voice recognition program, and a computer-readable non-transitory recording in which the program is stored. It can be considered as a medium, a voice recognition system, a device control device, a vehicle, or the like.

  INDUSTRIAL APPLICABILITY The present invention is useful for a voice recognition device mounted on a vehicle or the like.

10 voice recognition device 11 voice recognition unit 12 scene identification unit 13 storage unit 14 device control unit 20 microphone 31 first device 32 second device 33 second device

Claims (1)

A voice recognition device capable of controlling a device according to an input voice,
A voice recognition unit that accepts voice input, performs voice recognition, and specifies the utterance content,
Acquiring the state of the device, as a current scene a scene determined based on the state, a scene specifying unit for specifying from a plurality of scenes,
A storage unit that stores in advance one or more combinations of utterance content and control content in association with each of the plurality of scenes,
If the utterance content identified by the voice recognition unit is associated with the current scene by referring to the storage unit, the utterance content identified by the voice recognition unit is associated with the current scene. A voice recognition device, comprising: a device control unit that executes combined control contents.

Images