Voice operating system, server device, on-vehicle device, and voice operating method
A purpose of the present invention is to provide a technique for easily performing accurate voice recognition. A voice operating system includes an on-vehicle device and a server device, in which the on-vehicle device includes a voice-input accepting unit, a vehicle-environment-information management unit, and an output control unit that transmits input information and vehicle-environment information to the server device, then outputs an execution task candidate received from the server device together with an estimated misrecognition factor, and transmits, when accepting permission, an instruction to execute the execution task candidate, and the server device includes a voice-data processing unit that converts the input information into a text, a task estimation unit that specifies the execution task candidate, reliability and the estimated misrecognition factor, a user-confirmation-necessity determination unit that transmits, when the reliability is less than a predetermined value, the estimated factor and the execution task candidate, and a processing unit.
1. A voice operating system comprising: an on-vehicle device; and a server device capable of communicating with the on-vehicle device, wherein
the on-vehicle device comprises:
a voice-input accepting unit configured to accept input information in voice;
a vehicle-environment-information management unit configured to acquire vehicle-environment information specifying an environment of a vehicle provided with the on-vehicle device; and
an output control unit configured to transmit the input information and the vehicle-environment information to the server device, then output an execution task candidate received from the server device together with an estimated misrecognition factor of voice recognition, and transmit, when accepting execution permission of the execution task candidate, an instruction to execute the execution task candidate to the server device, and
the server device comprises:
a voice-data processing unit configured to convert, when accepting the input information and the vehicle-environment information from the on-vehicle device, the input information into a character string;
a task estimation unit configured to specify the execution task candidate corresponding to the character string into which the voice-data processing unit converts the input information, reliability using the input information and the vehicle-environment information, and the estimated misrecognition factor if any;
a user-confirmation-necessity determination unit configured to transmit, when the reliability is less than a predetermined value, the estimated misrecognition factor of the voice recognition and the execution task candidate to the on-vehicle device; and
a task processing unit configured to execute the execution task candidate;
wherein
the on-vehicle device further comprises
a task-estimation-result-history management unit configured to manage a history of the execution task candidate,
the task-estimation-result-history management unit stores the execution task candidate received by the output control unit, and
the task estimation unit excludes
the execution task candidate stored in the task-estimation-result-history management unit in processing for specifying the execution task candidate.
2. The voice operating system according to claim 1 , wherein
the output control unit notifies in voice, when accepting a re-input instruction for the execution task candidate, a coping method for the estimated misrecognition factor of the voice recognition, and causes the voice-input accepting unit to accept input information in voice.
3. The voice operating system according to claim 1 , wherein
the on-vehicle device comprises
a reliability-decreasing-factor storage unit configured to store a history of the estimated misrecognition factor of the voice recognition received by the output control unit,
the task-estimation-result-history management unit stores the estimated misrecognition factor of the voice recognition received by the output control unit, and
the task estimation unit excludes
the estimated misrecognition factor of the voice recognition stored in the reliability-decreasing-factor storage unit in processing for specifying the estimated misrecognition factor.
4. The voice operating system according to claim 1 , wherein
the output control unit notifies in voice, when accepting a re-input instruction for the execution task candidate, a coping method for the estimated misrecognition factor of the voice recognition, and causes the voice-input accepting unit to accept input information in voice, and notifies, when an utterance sound pressure of the input information in voice previously accepted by the voice-input accepting unit is smaller than a predetermined sound pressure, the coping method by increasing a volume to be equal to or larger than a predetermined sound volume in the processing.
5. The voice operating system according to claim 1 , wherein
the output control unit notifies in voice, when accepting a re-input instruction for the execution task candidate, a coping method for the estimated misrecognition factor of the voice recognition, and causes the voice-input accepting unit to accept input information in voice, and stops, when a noise sound pressure of the input information in voice previously accepted by the voice-input accepting unit is larger than a predetermined sound pressure, accepting the input information in voice in the processing.
6. A server device capable of communicating with a predetermined on-vehicle device, the server device comprising:
a voice-data processing unit configured to convert, when accepting input information in voice and vehicle-environment information specifying an environment of a vehicle provided with the on-vehicle device from the on-vehicle device, the input information into a character string;
a task estimation unit configured to specify an execution task candidate corresponding to the character string into which the voice-data processing unit converts the input information, reliability using the input information and the vehicle-environment information, and an estimated misrecognition factor if any;
a user-confirmation-necessity determination unit configured to transmit, when the reliability is less than a predetermined value, the estimated misrecognition factor of voice recognition and the execution task candidate to the on-vehicle device; and
a task processing unit configured to execute the execution task candidate,
wherein the task estimation unit inquires of the on-vehicle device a history of the execution task candidate received by the on-vehicle device and excludes the execution task candidate of the history in processing for specifying the execution task candidate.
7. A voice operating method for a voice operating system comprising an on-vehicle device and a server device capable of communicating with the on-vehicle device, wherein
the on-vehicle device comprises a control unit,
the control unit executes:
a voice-input accepting step of accepting input information in voice;
a vehicle-environment-information management step of acquiring vehicle-environment information specifying an environment of a vehicle provided with the on-vehicle device; and
an output control step of transmitting the input information and the vehicle-environment information to the server device, then outputting an execution task candidate received from the server device together with an estimated misrecognition factor of voice recognition, and transmitting, when accepting execution permission of the execution task candidate, an instruction to execute the execution task candidate to the server device,
the server device comprises a server control unit, and
the server control unit executes:
a voice-data processing step of converting, when accepting the input information and the vehicle-environment information from the on-vehicle device, the input information into a character string;
a task estimation step of specifying the execution task candidate corresponding to the character string into which the input information is converted in the voice-data processing step, reliability using the input information and the vehicle-environment information, and the estimated misrecognition factor if any;
a user-confirmation-necessity determination step of transmitting, when the reliability is less than a predetermined value, the estimated misrecognition factor of the voice recognition and the execution task candidate to the on-vehicle device; and
a task processing step of executing the execution task candidate;
wherein
the on-vehicle device comprises
a task-estimation-result-history management unit configured to manage a history of the execution task candidate,
the task-estimation-result-history management unit stores the execution task candidate received by the output control unit, and
the task estimation unit excludes
the execution task candidate stored in the task-estimation-result-history management unit in processing for specifying the execution task candidate.
8. A voice operating system comprising: an on-vehicle device; and a server device capable of communicating with the on-vehicle device, wherein
the on-vehicle device comprises:
a voice-input accepting unit configured to accept input information in voice;
a vehicle-environment-information management unit configured to acquire vehicle-environment information specifying an environment of a vehicle provided with the on-vehicle device; and
an output control unit configured to transmit the input information and the vehicle-environment information to the server device, then output an execution task candidate received from the server device together with an estimated misrecognition factor of voice recognition, and transmit, when accepting execution permission of the execution task candidate, an instruction to execute the execution task candidate to the server device, and
the server device comprises:
a voice-data processing unit configured to convert, when accepting the input information and the vehicle-environment information from the on-vehicle device, the input information into a character string;
a task estimation unit configured to specify the execution task candidate corresponding to the character string into which the voice-data processing unit converts the input information, reliability using the input information and the vehicle-environment information, and the estimated misrecognition factor if any;
a user-confirmation-necessity determination unit configured to transmit, when the reliability is less than a predetermined value, the estimated misrecognition factor of the voice recognition and the execution task candidate to the on-vehicle device; and
a task processing unit configured to execute the execution task candidate,
wherein
the on-vehicle device comprises
a reliability-decreasing-factor storage unit configured to store a history of the estimated misrecognition factor of the voice recognition received by the output control unit,
the task-estimation-result-history management unit stores the estimated misrecognition factor of the voice recognition received by the output control unit, and
the task estimation unit excludes
the estimated misrecognition factor of the voice recognition stored in the reliability-decreasing-factor storage unit in processing for specifying the estimated misrecognition factor.