Agent device, system, control method of agent device, and storage medium

A control method of an agent device comprising:

• by a computer,
• imaging an occupant of the vehicle;
• detecting the occupant of the vehicle;
• providing a service which includes causing an output device to output a voice response according to a voice of the occupant which is collected in a vehicle interior of a vehicle;
• receiving an input by the occupant;
• changing or adding a starting method of a function of providing the service on the basis of the received content;
• acquiring an image in response to receiving a first starting keyword, the first starting keyword being a word that is set by the first occupant and for causing one or more agent controllers to start;
• referring to a first reference information to determine whether a feature of the first occupant extracted from the image matches a feature associated with identification information of the first occupant in the first reference information or not, the first reference information being the feature of the first occupant extracted from an image and the identification information of the first occupant are associated with each other;
• determining that the first occupant has input the first starting keyword when the feature of the first occupant extracted from the image is determined to match the feature associated with identification information of the first occupant;
• determining that the first occupant has not input the first starting keyword when the feature of the first occupant extracted from the image is determined to not match the feature associated with identification information of the first occupant;
• not starting the one or more agent controllers in accordance with the first starting keyword in response to receiving the first starting keyword and determining that the first occupant has not input the first starting keyword;
• starting the one or more agent controllers in accordance with the first starting keyword in response to receiving the first starting keyword and determining that the first occupant has input the first starting keyword;
• providing the service with reference to a first operation history, when the one or more agent controllers start in accordance with the first starting keyword, the first operation history being operation history in which the one or more agent controllers provided the service in past when the agent controller has started with the first starting keyword;
• starting the one or more agent controllers in accordance with a second starting keyword that is different form the first starting keyword regardless of a result of the detection of the occupant detector when the receiver receives the second starting keyword;
• and
• providing the service with reference to a second operation history that is different form the first operation history when the one or more agent controllers start in accordance with the second starting keyword.

A non-transitory computer-readable storage medium that stores a computer program to be executed by a computer to perform at least:

• image an occupant of the vehicle;
• detect the occupant of the vehicle;
• provide a service which includes causing an output device to output a voice response according to a voice of the occupant which is collected in a vehicle interior of a vehicle;
• receive an input by the occupant;
• change or add a starting method of a function of providing the service on the basis of the received content;
• acquire an image in response to receiving a first starting keyword, the first starting keyword being a word that is set by the first occupant and for causing one or more agent controllers to start;
• refer to a first reference information to determine whether a feature of the first occupant extracted from the image matches a feature associated with identification information of the first occupant in the first reference information or not, the first reference information being the feature of the first occupant extracted from an image and the identification information of the first occupant are associated with each other;
• determine that the first occupant has input the first starting keyword when the feature of the first occupant extracted from the image is determined to match the feature associated with identification information of the first occupant;
• determine that the first occupant has not input the first starting keyword when the feature of the first occupant extracted from the image is determined to not match the feature associated with identification information of the first occupant;
• not start the one or more agent controllers in accordance with the first starting keyword in response to receiving the first starting keyword and determining that the first occupant has not input the first starting keyword;
• start the one or more agent controllers in accordance with the first starting keyword in response to receiving the first starting keyword and determining that the first occupant has input the first starting keyword;
• provide the service with reference to a first operation history, when the one or more agent controllers start in accordance with the first starting keyword, the first operation history being operation history in which the one or more agent controllers provided the service in past when the agent controller has started with the first starting keyword;
• start the one or more agent controllers in accordance with a second starting keyword that is different form the first starting keyword regardless of a result of the detection of the occupant detector when the receiver receives the second starting keyword;
• and
• provide the service with reference to a second operation history that is different form the first operation history when the one or more agent controllers start in accordance with the second starting keyword.