WO2019064928A1 - 情報処理装置、情報処理システム、及び情報処理方法 - Google Patents

情報処理装置、情報処理システム、及び情報処理方法 Download PDF

Info

Publication number
WO2019064928A1
WO2019064928A1 PCT/JP2018/029218 JP2018029218W WO2019064928A1 WO 2019064928 A1 WO2019064928 A1 WO 2019064928A1 JP 2018029218 W JP2018029218 W JP 2018029218W WO 2019064928 A1 WO2019064928 A1 WO 2019064928A1
Authority
WO
WIPO (PCT)
Prior art keywords
text
vehicle
unit
general
information processing
Prior art date
Application number
PCT/JP2018/029218
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
岡本　圭司
Original Assignee
株式会社デンソー
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社デンソー filed Critical 株式会社デンソー
Publication of WO2019064928A1 publication Critical patent/WO2019064928A1/ja

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60R—VEHICLES, VEHICLE FITTINGS, OR VEHICLE PARTS, NOT OTHERWISE PROVIDED FOR
    • B60R16/00—Electric or fluid circuits specially adapted for vehicles and not otherwise provided for; Arrangement of elements of electric or fluid circuits specially adapted for vehicles and not otherwise provided for
    • B60R16/02—Electric or fluid circuits specially adapted for vehicles and not otherwise provided for; Arrangement of elements of electric or fluid circuits specially adapted for vehicles and not otherwise provided for electric constitutive elements
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/40—Processing or translation of natural language
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue

Definitions

  • the present disclosure relates to an information processing apparatus, an information processing system, and an information processing method, and more particularly to an information processing apparatus for a vehicle using an interactive interface, an information processing system, and an information processing method.
  • a messaging platform which realizes chat (chat), various messengers (Messenger), etc. (hereinafter referred to as "messaging") in a portable terminal in which conversation is performed via voice or text (character) of natural language. Platform) has become widespread.
  • an interactive agent has been developed in which various results are obtained by conducting conversations with a user using an interactive interface similar to messaging.
  • Siri manufactured by Apple Inc.
  • an interactive agent of the iPhone registered trademark
  • Patent Document 1 as a conventional interactive agent for vehicles, the driving situation is observed based on the sensor information, and the observation result is learned by storing it together with the sensor information, and based on the learning result
  • a vehicle agent device is described which determines an act of communication with a user and displays an agent representing the determined act of communication.
  • the agent apparatus for vehicles of patent documents 1 makes the learning result reflect information acquired outside the vehicle accumulated in a portable terminal unit, and determines a communication act.
  • AI Artificial Intelligence
  • An object of the present disclosure is to provide an information processing apparatus, an information processing system, and an information processing method.
  • an information processing apparatus includes a state acquisition setting unit that acquires vehicle data including a state of a vehicle, and a conversation for a vehicle corresponding to the vehicle data acquired by the state acquisition setting unit.
  • General-purpose text for creating general-purpose conversation text which is text for general-purpose conversation that is different from the text for vehicle that is created by the text-for-vehicle for vehicle that creates text for vehicle that is text
  • a text selection unit for selecting the text for the vehicle created by the text creation unit for the vehicle according to a specific judgment criterion and the general-purpose text created by the general-purpose text creation unit according to a specific judgment standard;
  • a text output unit for outputting the vehicle text selected by the unit or the general-purpose text.
  • general-purpose text which is text of a general-purpose conversation different from the text for a vehicle is created, and the text for a vehicle and the general-purpose are By selecting and outputting the text, it is possible to appropriately respond to questions other than the question concerning the vehicle, and it is possible to provide an information processing apparatus capable of making the user feel familiar.
  • an information processing system includes an information processing device and a terminal connected to the information processing device.
  • the information processing apparatus includes a state acquisition setting unit that acquires vehicle data including a state of a vehicle, a text acquisition unit that acquires an inquiry text by a user from the terminal, and the vehicle data acquired by the state acquisition setting unit.
  • a vehicle text creation unit that creates a text for a vehicle that is text of a conversation corresponding to at least one of the questioning texts acquired by the text acquisition unit, and calculates a degree of matching of the created text for a vehicle
  • a general-purpose text which is text of a general-purpose conversation corresponding to the questioning text acquired by the text acquisition unit, which is different from the vehicle text generated by the vehicle text generation unit
  • a general-purpose text creation unit that calculates the degree of matching of the general-purpose text, and the vehicle according to a specific determination criterion
  • a text selection unit for selecting the vehicle text and the general-purpose text based on the degree of conformity of the text for vehicle generated by the text-for-text generation unit and the degree of conformity of the general-purpose text generated by the general-purpose text generation unit
  • a text output unit for outputting the vehicle text selected by the text selection unit or the general-purpose text to the terminal.
  • the terminal includes a text transmission unit that transmits the inquiry text to the information processing apparatus, and a text
  • general-purpose text which is text of a general-purpose conversation different from the text for a vehicle is created, and the text for a vehicle and the general-purpose are By selecting and outputting the text, it is possible to appropriately respond to questions other than the question concerning the vehicle, and it is possible to provide an information processing apparatus capable of making the user feel familiar.
  • an information processing method executed by an information processing apparatus acquires vehicle data including a state of a vehicle, and the vehicle is a conversation text for a vehicle corresponding to the acquired vehicle data.
  • Texts for general purpose, texts for general conversations other than the texts for vehicles, and general texts for texts, and the text for vehicles created according to a specific judgment criterion and the general texts created And outputting the selected text for vehicle or the general-purpose text.
  • general-purpose text which is text of a general-purpose conversation different from the text for a vehicle is created, and the text for a vehicle and the general-purpose are By selecting and outputting the text, it is possible to appropriately respond to questions other than the question concerning the vehicle, and it is possible to provide an information processing apparatus capable of making the user feel familiar.
  • FIG. 1 is a system configuration diagram of an information processing system according to a first embodiment of the present disclosure
  • FIG. 2 is a block diagram showing a control configuration of the text input / output device shown in FIG. It is a conceptual diagram of response DB of the conversation server for vehicles shown in FIG. It is a block diagram which shows the control structure of the terminal shown in FIG. 5 is a flowchart of text input / output processing according to the first embodiment of the present disclosure
  • FIG. 6 is a flowchart of (a) vehicle data storage processing and (b) vehicle text creation processing shown in FIG.
  • FIG. 5 It is a screen example of text input / output processing shown in FIG. It is a block diagram showing composition of an information processing system concerning a second embodiment of the present disclosure, It is a block diagram showing composition of an information processing system concerning a third embodiment of this indication, It is a block diagram showing composition of an information processing system concerning a 4th embodiment of this indication, It is a block diagram showing composition of an information processing system concerning a 5th embodiment of this indication.
  • the information processing system Xa is a system that performs control for a vehicle using an interactive interface.
  • the information processing system Xa also includes a text input / output device 1a, a car 2, and a terminal 3a.
  • the text input / output device 1a is an information processing device connected to the terminal 3a and including various servers.
  • the text input / output device 1a implements an interactive agent that inputs / outputs text in an interactive interface by messaging.
  • the text input / output device 1a includes a vehicular conversation server 10a, a general conversation server 11a, and a messaging server 12a.
  • an information processing apparatus including other servers and the like according to another embodiment is simply referred to as the text input / output apparatus 1 when collectively shown.
  • the vehicular conversation server 10a creates a text of the conversation for the vehicle. Specifically, the vehicular conversation server 10a creates a vehicular text 400 (FIG. 2) for response corresponding to the state of the vehicle. For this reason, the conversation server for vehicles 10a can use an AI agent specialized for vehicles.
  • the vehicular conversation server 10a acquires the general-purpose text 410 from the general-purpose conversation server 11a, selects one that is determined to be appropriate from the vehicular text 400, and returns a response to the user via the messaging server 12a.
  • This response is, for example, a statement of the personified car 2.
  • the general-purpose conversation server 11a creates general-purpose text 410 that is text of a general-purpose conversation. Therefore, the general-purpose conversation server 11a can use a general-purpose conversation AI agent.
  • the messaging server 12a is a server of a messaging platform that provides a service of messaging.
  • the messaging server 12a provides an interactive interface for inputting and outputting natural language sentences (texts), and transmits a user's inquiry from the terminal 3a to the vehicular conversation server 10a.
  • the vehicular conversation server 10a and / or the messaging server 12a may, for example, distribute or install the application to the terminal 3a, manage the operation or setting of the application, or acquire various information. . Also, dedicated servers for these may be prepared.
  • the conversation server 10a for vehicles, the general purpose conversation server 11a, and the messaging server 12a may be provided with two or more.
  • each of these servers may also include a plurality of servers.
  • the terminal 3a is an information processing apparatus such as a smart phone (Smart Phone), a mobile phone, a PDA (Personal Data Assistant), a PC (Personal Computer), a home appliance, a car, a motorcycle, and the like.
  • the terminal 3a has a messaging function. Therefore, the terminal 3a can transmit and receive text with the interactive input / output device with the text input / output device 1a.
  • the terminal 3a may be provided with various sensors and reception units such as GPS.
  • the terminal 3a may be owned or lent by a user who may be an owner, a driver, a passenger or the like of the automobile 2.
  • an information processing apparatus including portable terminals and the like having another configuration according to another embodiment is simply referred to as a terminal 3 when collectively shown.
  • the automobile 2 is a so-called "connected car” or the like, and is a vehicle of the present embodiment.
  • the car 2 transmits various states to the text input / output device 1a. Further, the automobile 2 can also perform various settings and controls by a command or the like from the text input / output device 1a.
  • the car 2 can also detect whether the user is in the car or in operation by communication with a seating sensor or the terminal 3a. This detection may be configured to detect a sporadic with a beacon or the like.
  • each server of the text input device a is a cellular phone network, WAN (Wide Area Network) such as Internet (registered trademark), or Wifi (registered trademark), wireless LAN, etc. It may be connected by an IP network such as LAN (Local Area Network).
  • WAN Wide Area Network
  • IP network such as LAN (Local Area Network).
  • each device can acquire and synchronize time from GPS, a base station, a time server, or the like.
  • the vehicle conversation server 10a, the general conversation server 11a, and the messaging server 12a of the text input / output device 1a include control units 100, 101, 102, storage units 110, 111, 112, and the like, respectively.
  • the respective units of each server are connected to the control units 100, 101, 102, respectively, and are operation controlled by the control units 100, 101, 102.
  • the control units 100, 101, 102 are CPU (Central Processing Unit, Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), DSP (Digital Signal Processor), other ASIC (Application Specific Processor), Application processor etc.
  • CPU Central Processing Unit
  • MPU Micro Processing Unit
  • GPU Graphics Processing Unit
  • DSP Digital Signal Processor
  • ASIC Application Specific Processor
  • each control unit 100, 101, 102 reads out the control program stored in the auxiliary storage unit of each corresponding storage unit 110, 111, 112, develops this control program in the main storage unit, and executes it. Thus, it can be operated as each functional block (functional unit, component) described later. Further, the control units 100, 101, 102 respectively control the whole of the vehicular conversation server 10a, the general conversation server 11a, and the messaging server 12a.
  • the storage units 110, 111, and 112 are non-transitory tangible storage media. Further, the storage units 110, 111, 112 are a main storage unit such as a random access memory (RAM), an auxiliary storage unit such as a read only memory (ROM), a solid state disk (SSD), a hard disk drive (HDD), etc. A memory card, an optical recording medium, etc. may be included.
  • RAM random access memory
  • ROM read only memory
  • SSD solid state disk
  • HDD hard disk drive
  • a memory card, an optical recording medium, etc. may be included.
  • control programs for performing operation control of the vehicle conversation server 10a, the general conversation server 11a, and the messaging server 12a are stored, respectively.
  • the control program may store an operating system (OS), a server program (daemon), an application, and the like.
  • OS operating system
  • server program daemon
  • application application
  • storage units 110, 111, 112 also store various data.
  • Each server may further include a connection unit including a wireless transceiver for connecting to a network, a LAN interface, and the like.
  • control unit 100 of the vehicular conversation server 10a includes a state acquisition setting unit 200, a vehicle text creation unit 220, and a text selection unit 240.
  • the storage unit 110 of the vehicular conversation server 10a stores vehicle data 300, inquiry text 310, response DB 320, judgment reference data 330, text for vehicle 400, general-purpose text 410, and degree of matching 401, 411.
  • control unit 101 of the general-purpose conversation server 11 a includes a general-purpose text creation unit 230.
  • the storage unit 111 of the general-purpose conversation server 11a stores an inquiry text 310, a response DB 321, a general-purpose text 410, and a matching degree 411.
  • control unit 102 of the messaging server 12 a includes a text acquisition unit 210 and a text output unit 250.
  • the storage unit 112 of the messaging server 12a stores the inquiry text 310, the selection text 420, and the schedule data 421.
  • the state acquisition setting unit 200 acquires vehicle data 300 including the state of the vehicle. Specifically, the state acquisition setting unit 200 acquires the vehicle data 300 from the automobile 2 and stores the acquired vehicle data 300 in the storage unit 110. At this time, the state acquisition setting unit 200 may optimize communication by accumulating cache (temporary data).
  • the state acquisition setting unit 200 can also perform setting and control of the automobile 2.
  • the state acquisition setting unit 200 may perform setting or control of the automobile 2 according to a command instructed by the user.
  • This command may be acquired by interpreting the question text 310 in the vehicular text creation unit 220 described later.
  • this command may include an instruction to set each part of the car 2, to set or acquire an image of an on-vehicle camera, to set navigation, to perform autopilot, to acquire vehicle data 300, and the like.
  • the text acquisition unit 210 acquires the inquiry text 310 input by the user from the terminal 3a.
  • the text acquisition unit 210 executes processing for receiving various messages of the messaging platform, receives the inquiry text 310 by WWW (World Wide Web) or a unique protocol, and temporarily stores it in the storage unit 112. Then, the text acquisition unit 210 transmits the acquired inquiry text 310 to the vehicle text creation unit 220 and the general-purpose text creation unit 230.
  • the text acquisition unit 210 may transmit and store the inquiry text 310 to the storage unit 110 of the vehicular conversation server 10a and the storage unit 111 of the general-purpose conversation server 11a.
  • the vehicular text creation unit 220 creates the vehicular text 400 corresponding to the vehicle data 300 acquired by the state acquisition setting unit 200 and / or the inquiry text 310 acquired by the text acquisition unit 210, and stores the vehicle text 400 in the storage unit 110. Do.
  • the text-for-vehicle creating unit 220 can use an AI agent for chat designed for a specific area.
  • this "specific region" may be for a vehicle. That is, the text-for-vehicle creating unit 220 may be an AI agent that enables conversation specific to the car 2.
  • the vehicular text creation unit 220 creates, for example, a vehicular text 400 that is text of a conversation for a vehicle. That is, the vehicle text creation unit 220 creates a vehicle text 400 specialized for the personified car 2.
  • the text-for-vehicle creating unit 220 analyzes the state of the vehicle from the vehicle data 300 in order to create the text for vehicle 400. Further, the text-for-vehicle creating unit 220 performs analysis by natural language processing such as keyword extraction, morphological analysis, corpus analysis and the like from the inquiry text 310. In addition, the text-for-vehicle creating unit 220 determines responses using an AI engine including various AIs such as decision trees, artificial neural networks, kernel machines, reinforcement learning, Bayesian networks, hidden Markov models, and various statistical methods. , Natural language processing to create a text 400 for a vehicle.
  • AI engine including various AIs such as decision trees, artificial neural networks, kernel machines, reinforcement learning, Bayesian networks, hidden Markov models, and various statistical methods.
  • the text-for-vehicle creating unit 220 may create, for example, text for vehicle 400 for "voluntary" conversation starting from the personified AI agent side.
  • the text-for-vehicle creating unit 220 opens the window as “notice” to the user, does not lock the key, keeps the hazard lamp on while the headlight is on, reminds the user of the schedule, etc.
  • You may create the text 400 for vehicles which gives notice of. Among these, “schedule reminder” may be notified when the user has set navigation and the like in advance.
  • the vehicle text creation unit 220 notifies the user of a decrease in the remaining amount of gasoline, a decrease in tire air pressure, lighting of various warning lamps, an intruder warning, a security alarm and the like as a "warning" to the user. You may create 400. In addition, it may be inconvenient to notify these warnings immediately when an event occurs. For example, in the case of warning lighting of a hazard lamp or a headlight, it may be considered necessary to light the lamp in the first place. In such a case, the text-for-vehicle preparing unit 220 may cause the text selecting unit 240 described later to perform scheduling or the like so as to notify at a sufficient time interval.
  • the text-for-vehicle creating unit 220 receives, from the vehicle data 300, the receipt of the courier service via the trunk, the notification from the dealer, the advertisement based on the geographical information, etc. as related to various "additional functions". You may create the text 400 for vehicles which notifies. That is, the text-for-vehicle creating unit 220 may create a state for receiving a home delivery service, a notice such as a vehicle inspection time or a legal check from a dealer of the automobile 2, an advertisement based on geographical information, or the like as the text for vehicle 400. In this "state of receipt of the courier service", when the permission to use the trunk of the car 2 as a delivery box is set, the delivery company delivers the package and puts the package in the trunk of the vehicle 2 etc. Indicates
  • Vehicle text 400 may be created.
  • the vehicular text creation unit 220 may interpret a command instructed by the user from the inquiry text 310, transmit this command to the state acquisition setting unit 200, and perform setting or control of the automobile 2. Also, this command may be designated by a keyword or the like. Further, this command may be acquired as a tag or the like included in the inquiry text 310 by pressing the GUI button added to a part of the interactive interface of the terminal 3a. Also, the command may include a command similar to the command provided in the conventional dedicated operation application. The command may be acquired even if it is other than the output time set in the schedule data 421, when there is an inquiry by the user.
  • the vehicular text creation unit 220 may create the vehicular text 400 corresponding to the phoneme of the question text 310, the edit distance, and / or the anagram, etc., using the response DB 320.
  • the text-for-vehicle creating unit 220 can entertain the user by creating the text for vehicle 400 for interaction according to the information related to the car 2 itself and the current state of the car 2.
  • the text-for-vehicle creating unit 220 may create a text for vehicle 400 for "game” by interpreting a suggestion etc. of the question text 310 such as "a game that talks about a car”. .
  • the vehicular text creation unit 220 calculates the matching degree 401 of the created vehicular text 400 and associates it with the vehicular text 400.
  • the text-for-vehicle creating unit 220 calculates, for example, the response score of the AI agent reflecting the state of the vehicle or the like included in the vehicle data 300 as the matching degree 401.
  • the response score of this AI agent may be a value based on the degree of certainty for the question text 310.
  • the certainty factor is, for example, 0 (unnatural) to 1 (natural) as a value indicating whether the response to the inquiry text 310 is natural for the user's speech or the like. It may be shown.
  • the vehicular text creation unit 220 can extract keywords from the query text 310, and morphological analysis can be reliably performed to extract words and grasp the meaning of sentences, and response sentences related to the words and sentences If it is possible to create a certainty, it may be judged as "natural" and the degree of certainty is high.
  • the vehicular text creation unit 220 may calculate a score using various evaluation criteria as the degree of conformity 401.
  • the text-for-vehicle creating unit 220 changes the calculation of the degree of matching 401 of the text for vehicle 400 according to the accumulation of the acquired vehicle data 300 and / or the presence or absence of the inquiry text 310. Good. More specifically, the text-for-vehicle creating unit 220 may change the calculation of the degree of matching 401 by changing the score of the AI agent, for example, depending on whether the vehicle data 300 is stored as a cache. For example, the text-for-vehicle creating unit 220 may calculate the degree of relevance 401 of the text for vehicle 400 created with no state of the vehicle to be transmitted, that is, no cash or old. Conversely, the text-for-vehicle creating unit 220 has the state of the vehicle to be transmitted, that is, calculates the degree of conformity 401 of the text 400 for vehicle created with the text for vehicle 400 in a state where cache is accumulated. May be
  • the text-for-vehicle creating unit 220 may change the calculation of the matching degree 401 of the text for vehicle 400 according to the output time and the state of the vehicle included in the vehicle data 300.
  • the vehicular text creation unit 220 may set the score high, when it is necessary to schedule the created vehicular text 400 to respond immediately. That is, the text-for-vehicle creating unit 220 may calculate a high score when it is desired to output immediately in scheduling described later.
  • the text-for-vehicle creating unit 220 may set the degree of matching 401 of the text 400 for vehicle created by the text-for-vehicle creating unit 220 high without acquiring the inquiry text 310. That is, the text-for-vehicle creating unit 220 can not acquire the questioning text 310 and calculates a high score when adding a “voluntary” response according to the state of the vehicle data 300, and adds this to the matching degree 401. It is also good.
  • the text-for-vehicle creating unit 220 may calculate a score using an evaluation criterion corresponding to the state of various vehicles included in the vehicle data 300.
  • the text-for-vehicle creating unit 220 may change the calculation of the degree of compatibility 401 in accordance with the state of the vehicle such as the remaining amount of gasoline, the tire pressure, the travel distance, and the warning light. That is, in the analyzed state of the vehicle, when there is something to be transmitted, the text-for-vehicle creating unit 220 may change the calculation of the matching degree 401.
  • the text-for-vehicle creating unit 220 may calculate the degree of conformity 401 by using the response DB 320 as a score based on the phoneme, the edit distance, the agreement of the anagram, the degree of conformity 401, and the like.
  • the text-for-vehicle creating unit 220 may change the calculation of the matching degree 401, for example, according to the difference in the method of the response of the AI agent including the state of the vehicle. Specifically, for example, when the vehicle data creating unit 220 obtains the state of the vehicle indicating that the remaining amount of gasoline is decreasing from the vehicle data 300, it is necessary to perform the response of the remaining amount of gasoline. Even in the case of the vehicle text 400 created by using the phoneme, the editing distance, the anagram, or the like, the matching degree 401 may be changed to be calculated high.
  • the text-for-vehicle creating unit 220 is created using phoneme, editing distance, anagram, etc.
  • the degree of matching 401 may be changed to be lower for the vehicle text 400.
  • the vehicular text creating unit 220 may lower the matching degree 401 so as not to operate during driving. Also, conversely, the text-for-vehicle creating unit 220 may create the text for vehicle 400 such as "operation prohibited during driving" and the like, and may increase the matching degree 401.
  • the text-for-vehicle creating unit 220 may notify the text selection unit 240 of the state of the vehicle analyzed from the vehicle data 300 or the like for scheduling or the like described later.
  • the general-purpose text creation unit 230 is a text of a general-purpose conversation corresponding to the inquiry text 310 acquired by the text acquisition unit 210, which is different from the vehicle-use text 400 generated by the vehicle text generation unit 220.
  • the general-purpose text 410 is created and stored in the storage unit 111. Therefore, the general-purpose text creation unit 230 can use an AI agent including various AI such as chat AI, chat bot, chat dialog API, and various statistical methods, which are designed to be able to perform general dialog response. is there. That is, the general-purpose text creation unit 230 may be an AI that enables general chat. Note that the general-purpose text creation unit 230 may be provided so as to be accessible as a Web API of a server on a so-called "cloud”.
  • the general-purpose text creation unit 230 may create the general-purpose text 410 by acquiring response data corresponding to the phoneme, the editing distance, and / or the anagram of the question text 310 from the response DB 321. At this time, the general-purpose text creation unit 230 can also create the general-purpose text 410 for the "game” by interpreting the "game” execution proposal etc. of the question text 310 such as "let's play shiri”. .
  • the general-purpose text creation unit 230 calculates the matching degree 411 of the created general-purpose text 410 and associates it with the general-purpose text 410.
  • the degree of matching 411 is the same as a value indicating whether the response to the inquiry text 310 is natural for the user's utterance or the like among the scores calculated by the vehicular text creating unit 220. It is also good. In this case, the score may be shown from 0 (unnatural) to 1 (natural).
  • the general-purpose text creation unit 230 may be able to respond to general conversations, but may not necessarily be optimal as a response specialized to the personified car 2. For example, even if the general-purpose text creation unit 230 asks for the remaining amount of gasoline of the automobile 2, it may not be able to respond correctly. In this case, the general-purpose text creation unit 230 may calculate the score of a value close to “0”.
  • the AI agent of the general-purpose text creation unit 230 may calculate the score of the response according to the following evaluation criteria: For example, for text 310 question of "Hello”, if you create a generic text 410 as "Hello", because it is judged to be appropriate, the score is high. That is, since the certainty factor of the response is high, the general-purpose text creation unit 230 calculates a high score.
  • the text selection unit 240 selects either the vehicle text 400 created by the vehicle text creation unit 220 or the general-purpose text 410 created by the general-purpose text creation unit 230 according to a specific evaluation criterion.
  • the text selection unit 240 refers to the determination reference data 330 and causes the text output unit 250 to output one of the selected vehicle text 400 and the general-purpose text 410 as the selected text 420. That is, the text selection unit 240 can selectively use the vehicle text 400 and the general-purpose text 410.
  • the text selection unit 240 sets schedule data 421 in order to perform scheduling for specifying the time of output of the selected text 420.
  • the text selection unit 240 may select the text using the degree of matching 401 of the text 400 for the vehicle and the degree of matching 411 of the general-purpose text 410 as the specific determination criteria. That is, the text selection unit 240 selects one of the degree of matching 401 of the text for vehicle 400 and the degree of matching 411 of the general-purpose text 410. As a result, the text selection unit 240 can switch the output response text using the scores output by the AI agents of the vehicle text creation unit 220 and the general-purpose text creation unit 230.
  • the text selection unit 240 may select the higher one of the matching degree 401 of the text 400 for the vehicle and the matching degree 411 of the general-purpose text 410. Furthermore, the text selection unit 240 may set an arbitrary bias to each of the matching degrees 401 and 411, compare them, and select the higher one. As a result, regardless of the setting of the order described later, the text with the higher score of the AI agent is selected. However, since each AI agent may calculate the score on a separate scale, direct comparison is difficult. For this reason, it is preferable that the text selection unit 240 sets and adjusts the threshold value and the bias of the judgment reference data 330 appropriately and then compares them.
  • the text selection unit 240 may make a selection by referring to the determination reference data 330 and corresponding to the setting of the order of selecting one of the vehicle text 400 and the general-purpose text 410. In addition, the text selection unit 240 may refer to the determination reference data 330 and use the threshold 401 or the bias of the matching degree 411 of the text 400 for the vehicle and the matching degree 411 of the general text 410 in this selection.
  • the text selection unit 240 can set and select the following order.
  • the vehicle text 400 is given priority. That is, when the matching degree 401 of the text 400 for vehicle is higher than the threshold value, the text selecting unit 240 selects the text 400 for vehicle.
  • the text selection unit 240 selects the general-purpose text 410 otherwise. In this way, priority is given to transmission to the user such as the condition of the vehicle included in the vehicle data on the vehicle 2, setting and control of the vehicle 2, provision of knowledge and maintenance information on the vehicle 2, provision of knowledge on general vehicles in general, etc.
  • the text selection unit 240 selects the universal text 410 when the matching degree 411 of the universal text 410 is higher than a threshold. In other cases, the text selection unit 240 selects the text 400 for vehicle. By configuring in this way, it is possible to realize natural responses to general dialogues, and to realize friendly conversations that distract topics from stories related to the car 2 when it is difficult to cope with them.
  • the text selection part 240 may be comprised so that this text 400 for vehicles may be selected preferentially.
  • the text selection unit 240 can change the “quality rating” of the response depending on the threshold of the degree of matching 401 and the degree of matching 411 regardless of the setting of the above-described order in either (1) or (2). It is. For example, if the text selection unit 240 sets the order to (1), the threshold of the degree of relevance 401 of the text 400 for the vehicle is extremely raised, “personality that twists any topic and talks about the car 2” It is possible to change the gender rating. In addition, when the text selection unit 240 sets the order to (1), when the threshold of the fitness 411 of the general-purpose text 410 is extremely high, "a personality that casually talks about the car 2" It is possible to change the gender rating.
  • the text selection unit 240 can also change the threshold value of the determination reference data 330. Specifically, the text selection unit 240 may change the score threshold in accordance with the vehicle state of the vehicle data 300 analyzed by the vehicle text creation unit 220 and various other situations. In addition, the text selection unit 240 may change the setting of the order of the determination reference data 330 in the same manner.
  • the text selection unit 240 may adjust the setting of the threshold and the order of the judgment reference data 330 according to the context before and after the question text 310 and the text of the response. Specifically, when the text selection unit 240 does not obtain the inquiry text 310 and responds from the vehicle text creation unit 220, setting of the threshold value and the order so that the vehicle text 400 is easily selected for a specific time May be changed. In addition, the text selection unit 240 may change the settings of the threshold value and the order so that the last response to the vehicle text creation unit 220 and the general-purpose text creation unit 230 is prioritized.
  • the text selection unit 240 temporarily changes the setting of the threshold and the order of the determination reference data 330 when the text preparation unit for vehicle 220 or the general-purpose text preparation unit 230 interprets the question text 310 as “game”. May be For example, when the text for vehicle 400 of the “game to talk about cars” is created, the text selection unit 240 determines the setting of an extreme threshold or order that gives priority to the text creation unit for vehicle 220. The reference data 330 is set. In this way, the phoneme of the response DB 320, the editing distance, and the anagram may be used to respond to any question regarding the car 2.
  • the text selection unit 240 sets an extreme threshold or order that prioritizes the general-purpose text creation unit 230 for the judgment reference data 330 Setting may be set in the judgment reference data 330.
  • the text selection unit 240 may perform scheduling to set the time of output of the text for vehicle 400 or the general-purpose text 410 by the text output unit 250. Specifically, the text selection unit 240 sets, as schedule data 421, a time at which the selected text 420 is output from the text output unit 250 as the scheduling. As a result, the time at which the selected text 420 is output can be changed depending on the condition of the vehicle 2 and the like.
  • the text selection unit 240 may perform scheduling corresponding to the type of state of the vehicle.
  • the text selection unit 240 can also change the scheduling of the selected text 420 based on, for example, the result of analyzing the state of the vehicle from the vehicle data 300 accumulated by the vehicle text creation unit 220. In this case, the text selection unit 240 changes, for example, the time of output of the selected text 420 already set in the schedule data 421.
  • the text selection unit 240 determines that the remaining amount of gasoline does not change on a weekday when the usage of the car 2 is different between a weekday and a weekend and the traveling distance tends to be longer at the weekend. You may also change the scheduling by assessing the remaining gasoline daily. In this case, although the text selection unit 240 has a sufficient amount of gasoline remaining for evaluation on weekdays, it may be reevaluated on holidays and scheduling may be performed to output a selected text 420 prompting refueling at that timing. .
  • the text selection unit 240 may change the scheduling so as to output again if the state of the vehicle is not the same and is not canceled.
  • the output time may be changed corresponding to the time when the selected text 420 was output last time.
  • the text selection unit 240 may perform scheduling based on the threshold value of the evaluation value for evaluating the output timing in the threshold value of the text for vehicle 400 of the determination reference data 330, based on this. That is, the text selection unit 240 calculates the evaluation value of the text for vehicle 400 at a specific time interval, and performs scheduling such that it is judged that the condition to be output is exceeded when the threshold of the evaluation value is exceeded. It is also good.
  • the text selection unit 240 may perform scheduling in which a condition for outputting the vehicle 2 “does not output during traveling” is set. This makes it possible to restrict the interaction with the personified AI using the terminal 3a while the driver is driving.
  • the information for the seating sensor may be included in the vehicle data 300 where the user operating the terminal 3a is seated, and the vehicular text creating unit 220 may analyze it based on this.
  • the text selecting unit 240 does not pass for a long time, and when the user's activity time such as midnight is inappropriate , Etc., may be scheduled so as not to output or at an appropriate time.
  • the text selection unit 240 changes the scheduling so that the selected text 420 is output when the inquiry text 310 from the user can be acquired. Good. That is, it may be considered that the user can receive the notification, and the selected text 420 may be output. Thus, it is possible to output the selected text 420 at an arbitrary time by the instruction of the text selection unit 240.
  • the text output unit 250 outputs the vehicle text 400 or the general-purpose text 410 selected by the text selection unit 240. Specifically, the text output unit 250 refers to the schedule data 421 for the selected text 420 acquired by the text selection unit 240, and outputs the selected text 420 to the terminal 3a in response to scheduling for specifying the time of output.
  • Selection text 420 may be output.
  • the text output unit 250 may perform control to output the selected text 420 according to the threshold of the evaluation value.
  • the text output unit 250 may perform transmission processing of the scheduling platform and output it to the terminal 3a.
  • Vehicle data 300 is data of a vehicle including the state of the vehicle.
  • the vehicle data 300 is data corresponding to the state of the automobile 2 which is a vehicle of the present embodiment.
  • the vehicle data 300 indicates, as the state of the car 2, the operating state of the car 2 according to the state of the switch or key, whether or not it is traveling, traveling speed, whether the user or driver of the terminal 3a is driving Opening and closing conditions of each window, key lock status, hazard lamp status, headlight status, car navigation settings, gasoline remaining amount, tire pressure, running distance, status of various warning lights, intruder alert, security alarm etc.
  • Data of multiple attributes such as information may be included.
  • the vehicle data 300 may be acquired in real time (real time) or accumulated as a cache for data of each of these attributes.
  • the data of each attribute of the vehicle data 300 may include information of the time when the data is acquired.
  • the vehicle data 300 may store the time etc. in a format such as a log (history).
  • vehicle data 300 may include, as an additional function, data such as reception of a courier service via a trunk, a notification from a dealer such as a vehicle inspection time, and an advertisement based on geographical information.
  • the response DBs 320 and 321 are response databases that accumulate response data for the inquiry text 310.
  • the determination criterion data 330 is data including a specific determination criterion when the text selection unit 240 selects the vehicle text 400 and the general-purpose text 410.
  • the determination criterion data 330 may include, as a particular determination criterion, the setting of the order of selecting one of the vehicle text 400 and the general-purpose text 410.
  • the criterion data 330 may include setting of a threshold value for selecting the matching degree 401 of the text 400 for the vehicle and the matching degree 411 of the general text 410.
  • the threshold may include the threshold, the bias, and the like of the matching degrees 401 and 411 when selecting the vehicle text 400 and the general-purpose text 410.
  • the inquiry text 310 is the text input by the user through messaging at the terminal 3a.
  • the inquiry text 310 may also include commands for setting and controlling the car 2. Further, the inquiry text 310 may include a command by a GUI button, for example, by a tag or the like.
  • the inquiry text 310 may not be simple text data, but may be encoded in units of words or characters.
  • the question text 310 may include emotion data and the like acquired at the time of voice input and at a sensor.
  • the vehicle text 400 is a text of a conversation corresponding to the vehicle data 300 acquired by the state acquisition setting unit 200 and / or the inquiry text 310 acquired by the text acquisition unit 210, generated by the vehicle text generation unit 220. is there.
  • the degree of matching 401 which is a score or the like of the AI agent of the text-for-vehicle creating unit 220, is added to the text for vehicle 400. That is, the degree of matching 401 is the degree of matching of the text 400 for vehicle.
  • the general-purpose text 410 is a text of a conversation corresponding to the inquiry text 310 acquired by the text acquisition unit 210 and created by the general-purpose dext preparation unit.
  • a matching degree 411 which is a score of an AI agent of the general-purpose dextrinsing unit, etc., is added. That is, the matching degree 411 is the matching degree of the general-purpose text 410.
  • the selection text 420 is the text 400 for vehicles or the general-purpose text 410 selected by the text selection unit 240.
  • Schedule data 421 which is a setting for scheduling, is added to the selection text 420.
  • the schedule data 421 may include data on time when the selected text 420 is output by the text output unit 250, setting of output conditions, a threshold of an evaluation value, and the like.
  • a plurality of selection texts 420 and schedule data 421 may be included.
  • the selection text 420 and the schedule data 421 may be stored in a stack or queue format.
  • the scheduled time may be different.
  • response database of this embodiment will be described with reference to FIG. 3 using the response DB 320 as a representative example.
  • the response DB 320 may be, for example, a database of a data structure in which vehicle knowledge, a vehicle element, an adjacent word dictionary, and a vehicle data attribute are associated.
  • the vehicle knowledge is, for example, know-how about the car 2 included in the text 400 for vehicle.
  • the vehicle element is, for example, an element related to a vehicle, a word corresponding to a keyword, or the like.
  • the word of the element related to the vehicle may correspond to, for example, one included in the word analyzed as the morpheme in the inquiry text 310.
  • the adjacent word dictionary may be a dictionary indicating words similar in phoneme, editing distance, anagram, and the like to words and the like analyzed from the query text 310.
  • the vehicle data attribute may indicate an attribute indicating the type of the state of the vehicle included in the vehicle data 300, which corresponds to the word in the above-described adjacent word dictionary.
  • the response DB 320 may further include response data such as various dictionaries and sentences.
  • control unit 100 of the vehicular conversation server 10a functions as the state acquisition setting unit 200, the vehicle text creation unit 220, and the text selection unit 240 by executing the control program stored in the storage unit 110. .
  • the control unit 101 of the general-purpose conversation server 11 a functions as a general-purpose text creation unit 230 by executing the control program stored in the storage unit 111.
  • the control unit 102 of the messaging server 12 a functions as a text acquisition unit 210 and a text output unit 250 by executing the control program stored in the storage unit 112.
  • the terminal 3a includes a control unit 103, a storage unit 113, an input unit 120, a display unit 130, and the like. Each unit is connected to the control unit 103, and operation control is performed by the control unit 103.
  • the control unit 103 includes a CPU, an MPU, a GPU, a DSP, and other ASICs.
  • the control unit 103 reads out and executes the control program of the storage unit 113 to operate as each functional block (functional unit) described later. Further, the control unit 103 controls the entire device of the terminal 3a.
  • the storage unit 113 is a non-transitional substantial recording medium.
  • the storage unit 113 may include a RAM, a ROM, an SSD, an HDD, a flash memory card, an optical recording medium, and the like.
  • the storage unit 113 stores a control program.
  • the control program includes an OS and an application.
  • the control program may also include a messaging application including chat, various messengers, a WWW browser, a terminal, etc. for realizing an interactive interface for connecting to a messaging platform.
  • the storage unit 113 also stores various data.
  • the input unit 120 is a touch panel, a pointing device such as a mouse, a button, a keyboard, an acceleration sensor, a gaze sensor, a biometric authentication sensor, a voice recognition device, or the like.
  • the input unit 120 acquires an input to the interactive interface or the like by the user, various instructions, and the like.
  • the text input by the input unit 120 is stored in the storage unit 113 as the inquiry text 310 via the OS and the application.
  • the input unit 120 may correspond to direct input of text by pressing a button such as a touch panel or a keyboard, direct input of a command by a GUI button, conversion of voice input or gesture input into text, or the like.
  • the display unit 130 is a liquid crystal display (LCD), an organic EL display, a light emitting diode (LED), or the like.
  • the inquiry text 310 On the display unit 130, the inquiry text 310, the selected text 420 received by the text receiving unit 260, and the like may be displayed by the messaging application.
  • the terminal 3a may also include an interface connected to a network such as a mobile phone network or a wireless LAN, Bluetooth (registered trademark), USB (Universal Serial Bus), various flash memory cards, SIM cards, and the like.
  • a network such as a mobile phone network or a wireless LAN, Bluetooth (registered trademark), USB (Universal Serial Bus), various flash memory cards, SIM cards, and the like.
  • control unit 103 of the terminal 3 a includes a text receiving unit 260 and a text transmitting unit 270.
  • the storage unit 113 stores the selected text 420 and the inquiry text 310.
  • the text transmission unit 270 transmits the inquiry text 310 stored in the storage unit 113 to the text input / output device 1a. At this time, the text transmission unit 270 may transmit the inquiry text 310 by HTTP, a proprietary protocol, or the like on the Internet via a mobile phone network, a wireless LAN, or the like.
  • the text transmission unit 270 transmits the inquiry text 310 to the text acquisition unit 210 of the messaging platform.
  • the text receiving unit 260 receives, from the text input / output device 1a, the selected text 420 which is the text 400 for vehicle or the general text 410.
  • the text receiving unit 260 receives the scheduled selected text 420 from the text output unit 250 of the messaging platform by HTTP, a proprietary protocol, or the like, and stores the selected text 420 in the storage unit 113.
  • the text receiving unit 260 may cause the display unit 130 to display the received selected text 420 using a messaging application.
  • the terminal 3a, the messaging server 12a, the vehicle conversation server 10a, and the general conversation server 11a may include other components in addition to the above-described units, and each unit may include a plurality of control units. .
  • any one of the components and any combination may be integrally configured.
  • the input unit 120 of the terminal 3a and the display unit 130 may be integrally formed.
  • each may be integrally formed also about the terminal 3a and the control part of each server, and a memory
  • control unit 103 of the terminal 3a is made to function as the text transmission unit 270 and the text reception unit 260 by executing a program such as an application such as messaging on the OS stored in the storage unit 113.
  • each part of the above-mentioned terminal 3a becomes a hardware resource which performs the method of this indication.
  • vehicle data 300 including the state of the vehicle is acquired.
  • a text for vehicle 400 is created.
  • general-purpose text 410 is created.
  • the created text for vehicle 400 and the created general-purpose text 410 are selected according to a specific determination criterion.
  • the selected text 420 which is the selected vehicle text 400 or the general text 410 is scheduled and output.
  • control units 100, 101, 102 of each server of the text input / output device 1a control the control programs stored in the storage units 110, 111, 112, the terminal 3a.
  • the unit 103 executes the control program stored in the storage unit 113 in cooperation with each unit using hardware resources.
  • Step S101 First, the state acquisition setting unit 200 of the text input / output device 1a performs vehicle data storage processing.
  • the state acquisition setting unit 200 acquires the vehicle data 300 from the automobile 2 and accumulates it in the storage unit 110 as a cache.
  • Step S301 Further, the control unit 103a of the terminal 3a performs an application activation process.
  • the terminal 3a starts and executes a messaging application when the device 3 is started up or restarted.
  • Step S302 Next, the text transmission unit 270 performs inquiry text transmission processing.
  • the query text 310 interactively input from the input unit 120 is stored in the storage unit 113 using the interactive interface of the messaging application.
  • the text transmission unit 270 transmits the stored inquiry text 310 to the text acquisition unit 210 of the text input / output device 1a.
  • Step S102 the text acquisition unit 210 of the text input / output device 1a performs an inquiry text acquisition process.
  • the text acquisition unit 210 acquires the inquiry text 310 transmitted from the text transmission unit 270 of the terminal 3a and temporarily stores it.
  • the text acquisition unit 210 transmits the stored inquiry text 310 to the vehicle text creation unit 220 and the general-purpose text creation unit 230.
  • the vehicular text creation unit 220 performs a vehicular text creation process.
  • the vehicular text creation unit 220 creates the vehicular text 400. Further, the vehicular text creation unit 220 calculates the degree of matching 401 of the created vehicular text 400 and adds the calculated degree 401 to the vehicular text 400.
  • Step S104 Next, the general-purpose text creation unit 230 performs general-purpose text creation processing.
  • the general-purpose text creation unit 230 creates general-purpose text 410.
  • the general-purpose text creation unit 230 calculates the degree of matching 411 of the created general-purpose text 410 and adds it to the general-purpose text 410.
  • the text selection unit 240 determines whether or not the general-purpose text 410 should be selected.
  • the text selection unit 240 refers to the determination criterion data 330 and makes a determination regarding selection of text to be responded according to a specific determination criterion. On this, the text selection unit 240 determines Yes when selecting the general-purpose text 410.
  • the text selection unit 240 determines No.
  • step S106 the text selection unit 240 proceeds with the process to step S106.
  • the text selection unit 240 proceeds with the process to step S107.
  • Step S106 When the general-purpose text 410 is selected, the text selection unit 240 performs general-purpose text selection processing.
  • the text selection unit 240 selects the general-purpose text 410 as the selection text 420. At this time, the text selection unit 240 may perform scheduling to specify the output time of the general-purpose text 410. In this case, the text selection unit 240 sets the schedule data 421 and adds it to the selected text 420.
  • Step S107 When the text for vehicle 400 is selected, the text selection unit 240 performs a text for vehicle selection process.
  • the text selection unit 240 selects the vehicle text 400 as the selection text 420. At this time, the text selection unit 240 may perform scheduling corresponding to the type of the state of the vehicle and / or scheduling specifying the output time of the text 400 for the vehicle. Also in this case, the text selection unit 240 sets the schedule data 421 and adds it to the selected text 420. (Step S108)
  • the text selection unit 240 determines whether or not the inquiry text 310 is acquired at a time other than the scheduled output time. In such a case, a situation such as a user's question instead of the "voluntary" statement output by the detailed process of the vehicle text creation process in step S103 shown in FIG. It is thought that. In such a case, it may be necessary to respond immediately. Therefore, the text selection unit 240 determines Yes when acquiring the inquiry text 310 at times other than the scheduled output time. In any other case, the text selection unit 240 determines No.
  • step S110 the text selection unit 240 proceeds with the process to step S110.
  • the text selection unit 240 proceeds with the process to step S109.
  • Step S109 the text selection unit 240 determines whether the selected text 420 should be output.
  • the text selection unit 240 refers to the schedule data 421 of the selected text 420 and determines Yes when the output condition such as the time for outputting the selected text 420 has come. If the text selection unit 240 does not output the selected text 420 yet, it determines No.
  • step S110 the text selection unit 240 proceeds with the process to step S110.
  • the text selection unit 240 ends the processing in the text input / output device 1a of the text input / output processing. (Step S110) When outputting the selected text 420, the text output unit 250 performs selected text output processing.
  • the text output unit 250 outputs the selected text 420 to the text receiving unit 260 of the terminal 3a.
  • the text reception unit 260 of the terminal 3a performs text reception processing.
  • the text receiving unit 260 receives the selected text 420 output from the text output unit 250 of the text input / output device 1a.
  • the text receiving unit 260 may respond to the user by displaying the received selected text 420 on the display unit 130 of the terminal 3a.
  • Step S121 the state acquisition setting unit 200 performs vehicle data acquisition processing.
  • the state acquisition setting unit 200 acquires the vehicle data 300 by, for example, transmitting data from the car 2 or inquiring the car 2 at time intervals such as several seconds to several hours. (Step S122) Next, the state acquisition setting unit 200 determines whether the vehicle data 300 can be acquired from the car 2 or not. Note that this process may be executed only when acquisition of the vehicle data 300 is attempted by an inquiry.
  • the state acquisition setting unit 200 determines that the result is Yes. In any other case, the state acquisition setting unit 200 determines No.
  • step S123 the state acquisition setting unit 200 proceeds with the process to step S123.
  • the state acquisition setting unit 200 ends the vehicle data accumulation process of the present embodiment. (Step S123) When acquisition of the vehicle data 300 is successful, the state acquisition setting unit 200 performs cache update processing.
  • State acquisition setting unit 200 accumulates vehicle data 300 as a cache.
  • the state acquisition setting unit 200 mainly caches, at an appropriate timing, attribute data that does not substantially change when the operation state of the vehicle 2 is not in operation in the attribute indicating the type of the state of the vehicle included in the vehicle data 300 It may be stored as For example, the state acquisition setting unit 200 may store the remaining amount of gasoline as cash because it does not change unless the gasoline is withdrawn by a pump or the like. In addition, the state acquisition setting unit 200 may accumulate, as a cache, an abnormality or the like that has occurred last. In addition, the status acquisition setting unit 200 does not resolve “oil level warning” or the like once it occurs, unless the oil is replenished. Therefore, the status acquisition setting unit 200 may store it as a cache in order to notify the user at necessary timing.
  • Step S131 the vehicular text creation unit 220 determines whether there is a vehicle state to be transmitted to the user. At this time, the text-for-vehicle creating unit 220 may determine whether there is a state of the vehicle to be transmitted by referring to the data of each attribute from the vehicle data 300 accumulated as the cache. If there is something to be transmitted in the state of the vehicle analyzed from the vehicle data 300, the vehicular text creation unit 220 determines as Yes. Otherwise, the vehicular text creation unit 220 determines No.
  • the text-for-vehicle creating unit 220 proceeds with the process to step S133.
  • step S132 the vehicle text creation unit 220 proceeds with the process to step S132.
  • Step S132 the text-for-vehicle preparing unit 220 determines whether the inquiry text 310 has been acquired. When the text for inquiry 310 is acquired, the vehicular text creation unit 220 determines that the result is Yes. Otherwise, the vehicular text creation unit 220 determines No.
  • the vehicular text creation unit 220 proceeds with the process to step S134.
  • the vehicle text creating unit 220 proceeds with the process to step S133.
  • Step S133 the vehicle text creation unit 220 performs a vehicle data creation process.
  • the vehicular text creation unit 220 creates a vehicular text 400 corresponding to the state of the vehicle analyzed from the vehicle data 300, and calculates and adds the matching degree 401.
  • the text-for-vehicle creating unit 220 may change the calculated degree of matching 401 to be higher according to the state of the vehicle.
  • the text for vehicle creation unit 220 is calculated even if the degree of relevance 401 is changed higher Good.
  • the text-for-vehicle creating unit 220 may notify the text selection unit 240 of the state of the vehicle.
  • the vehicular text creation unit 220 ends the vehicular text creation process according to the present embodiment.
  • the vehicle text creating unit 220 lowers the matching degree 401 according to the accumulation of the acquired vehicle data 300 and / or the presence or absence of the inquiry text 310. It is also possible to calculate by changing to. In this case, the degree of matching 401 may be calculated by the text selecting unit 240 in conjunction with the process of notifying the user of the warning or changing the scheduling according to the time of day. (Step S134) If the inquiry text 310 has been acquired, the vehicular text creation unit 220 performs response creation processing.
  • the vehicular text creation unit 220 creates a vehicular text 400 corresponding also to the inquiry text 310.
  • the vehicular text creation unit 220 analyzes the question text 310.
  • the text-for-vehicle creating unit 220 sends a command to the state acquisition setting unit 200 to set or control the vehicle 2.
  • the vehicular text creation unit 220 creates the vehicular text 400, and calculates and adds the matching degree 401 of the created vehicular text 400.
  • the vehicular text creation unit 220 may create the vehicular text 400 corresponding to the current state of the vehicle by the AI agent.
  • the text-for-vehicle creating unit 220 may create, for example, a text for vehicle 400 that responds to a question about the state of the vehicle while interacting with the user. At this time, it is also possible to include a warning rather than just a state.
  • the text-for-vehicle creating unit 220 transmits a command to turn off the hazard lamp to the state acquisition setting unit 200, thereby It is possible to check the status and create a vehicle text 400 such as "erased”.
  • the text-for-vehicle creating unit 220 can also create text for a vehicle 400 related to the automobile 2 itself to entertain the user.
  • the vehicular text creation unit 220 can create the following conversational sentence.
  • the text-for-vehicle creating unit 220 extracts the keyword of "Heart” for the question text 310 "I'm hungry", and "It's already time for dinner (T_T, too, the gasoline remaining amount is too It is possible to create a vehicle text 400 such as 1/10. ”.
  • the text-for-vehicle creating unit 220 may create the text for vehicle 400 that responds with a kind of “judgement” by using the phonology, the editing distance, the anagram, and the like with reference to the response DB 320.
  • the text-for-vehicle creating unit 220 morphologically analyzes the above-mentioned question text 310 "I'm hungry" and extracts morphemes. Then, the vehicular text creation unit 220 searches the adjacent word dictionary of the response DB 320 using the morpheme of the word “het”. Here, “tire-decrease-distance after tire replacement” is retrieved. The vehicle-text creating unit 220 may use these to create a vehicle-use text 400 "If it is reduced, my tires are also reduced.”.
  • the text-for-vehicle creating unit 220 also searches for “tire-dent-air pressure” of the response DB 320 from the morpheme of the word “dent” for the question text 310 “You are angry with your boss and get dented”. Ru. Using the above, the text-for-vehicle creating unit 220 uses the text for vehicle 400 saying "Don't get depressed! I'm about to dent my tires. You may create it.
  • the language “reduced, dented” expressing the user's emotion “frustrated, depressed” from the contents of the user's inquiry text 310 is the state of the car 2 " It is associated with the fact that fuel is running low and tires are worn out. Therefore, it is possible to entertain the user with an intelligent conversation that is consistent and can satisfy a certain level.
  • the text-for-vehicle creating unit 220 is not the cached vehicle data 300 but simply the information of the automobile 2 included in the vehicle knowledge of the response DB 320 Text 400 may be created.
  • the text-for-vehicle preparing unit 220 extracts “decreased-tire-replacement time” from the “heap” of the response DB 320 for the above-mentioned question text 310 “the hungry”.
  • the text-for-vehicle creating unit 220 creates the text for vehicle 400 that "If it is reduced, the tires will also be running. If it exceeds 30,000 km, please check the slip sign! You may
  • the text preparation part 220 for vehicles can also update response DB 320 by interaction with these users.
  • vehicle text creation unit 220 can create the vehicle text 400 "stitched" using a phoneme, an edit distance, an anagram, and the like.
  • the vehicular text creation unit 220 when using a phoneme, creates a vehicular text 400 that responds to so-called "rhyming" content. Specifically, for example, the text-for-vehicle creating unit 220 converts the morphologically analyzed word into roman letter reading and analyzes the phoneme. Then, the vehicular text creation unit 220 identifies similar phonological words from the dictionary or the like of the response DB 320, and creates a vehicular text 400 for conversation.
  • the text-for-vehicle creating unit 220 morphologically analyzes the question text 310 "puppies are cute" and makes the word of the morpheme “puppies” look like roman letters. That is, the "puppy” becomes “KOINU”. Then, since the two letters match with "OIRU", "oil” which is a phonological word is specified. As a result, the text-for-vehicle creating unit 220 can create the text for vehicle 400, such as “puppies and oil are similar! Please change the oil every one year or 2,000 km".
  • the vehicular text creation unit 220 may create the vehicular text 400 using an editing distance such as the Levenshtein distance.
  • the edit distance is the minimum number of steps required to transform one string into the other by inserting, deleting, and replacing each one character between the two strings. It may be a value. That is, the edit distance indicates how different the two character strings are.
  • the vehicular text creation unit 220 may use this to search a dictionary or the like of the response DB 320 as in the case of using a phoneme.
  • the vehicular text creation unit 220 may create the vehicular text 400 by selecting a word having a small edit distance and a high degree of phonetic match, using both the phoneme and the edit distance index.
  • the vehicular text creation unit 220 may create the vehicular text 400 using an anagram or the like.
  • the text for vehicle creation unit 220 may set the matching degree 401 high.
  • the conventional interactive interface unlike the conventional GUI, enables the humanization in which the user feels as if he / she is interacting with an object having a personality.
  • the information processing system Xa includes a text input / output device 1a, which is an information processing device, and a terminal 3a connected to the text input / output device 1a.
  • the text input / output device 1a is acquired by the state acquisition setting unit 200 acquiring vehicle data 300 indicating the state of the vehicle, the text acquisition unit 210 acquiring the inquiry text 310 by the user from the terminal 3a, and the state acquisition setting unit 200
  • Vehicle text 400 that is the text of a conversation corresponding to the question text 310 acquired by the acquired vehicle data 300 and / or the text acquisition unit 210 is created, and the vehicle 401 calculates the relevance 401 of the created vehicle text 400
  • General-purpose text that is general-purpose conversation text that corresponds to the inquiry text 310 acquired by the text acquisition unit 210, which is different from the vehicle text 400 generated by the for-use text generation unit 220 and the vehicle text generation unit 220 Create 410 and apply the generic text 410 created
  • the general-purpose text creation unit 230 that calculates the degree 411, the
  • the terminal 3a includes a text transmission unit 270 that transmits the inquiry text 310 to the text input / output device 1a, and a text reception unit 260 that receives the vehicle text 400 or the general-purpose text 410 from the text input / output device 1a.
  • the information processing system Xa can properly use the general-purpose text creation unit 230 and the vehicle-use text creation unit 220.
  • the vehicle text creation unit 220 reliably creates the vehicle text 400 regarding the communication related to the vehicle, in the weak question text 310
  • the general-purpose text creation unit 230 creates the general-purpose text 410, which is natural It is possible to make a response. Therefore, using the interactive interface, it is possible to continue to talk with the AI agent who personifies the vehicle without any particular purpose, and it is possible to have a pleasant dialogue. Thus, it is possible to help the user to attach to the car.
  • vehicle text creation unit 220 it is not necessary to provide the vehicle text creation unit 220 with perfect response performance, which facilitates development and leads to providing a system having an interactive interface that is useful as a whole.
  • the text input / output device 1a of the present embodiment is characterized in that the specific determination criterion is the degree of matching 401 of the text for vehicle 400 and the degree of matching 411 of the general-purpose text 410. That is, the text selection unit 240 can select any one of the degree of matching 401 of the text for vehicle 400 and the degree of matching 411 of the general-purpose text 410.
  • the specific determination criterion is the order of selecting one of the vehicle text 400 and the general-purpose text 410, and / or the matching degree 401 of the vehicle text 400 and the general-purpose text 410. It is characterized in that it also includes a threshold of the degree of matching 411 of
  • the text input / output device 1a of the present embodiment is characterized in that the text-for-vehicle creating unit 220 changes the calculation of the degree of matching 401 of the text for vehicle 400 according to the accumulation of the acquired vehicle data 300. I assume.
  • the text output unit 250 is selected text which is the text 400 for vehicle or the general-purpose text 410 even when the user makes an inquiry even if it is other than the output time. It is characterized by outputting 420.
  • the text input / output device 1a further includes response DBs 320 and 321 for storing response data to the inquiry text 310, and the vehicle text creating unit 220 and / or the general text creating unit 230 can respond to the response DB 320. , 321, to obtain response data corresponding to the phoneme of the question text 310, the edit distance, and / or the anagram, and to create the text of the conversation.
  • the text input / output device 1 a of the present embodiment is characterized in that the text selection unit 240 performs scheduling for setting the time of output of the text for vehicle 400 or the general-purpose text 410 by the text output unit 250.
  • a GUI button is also added to the messaging. This makes it possible to reliably transmit and execute the command. That is, by pressing the GUI button, text including commands in a fixed form is transmitted, and although it is an interactive interface, problems such as individual differences in writing do not occur.
  • general-purpose text 410 corresponding to this command can be created, and the variation of the conversation can be expanded.
  • the vehicle text creation unit 220 and the text selection unit 240 of the first embodiment are separated. That is, the vehicular conversation server 10b executes the vehicular text creation unit 220 exclusively.
  • the front end server 13 b has the function of the text selection unit 240.
  • the text of the user's query is sent to the front end server 13b.
  • the front end server 13b selects the text created by the general-purpose text creation unit 230 of the general-purpose conversation server 11b or the text-for-vehicle creation unit 220 of the conversation server for vehicle 10b according to the situation. That is, the front end server 13b responds to the user via the messaging platform of the messaging server 12b with text selected and obtained from different servers. This is the statement of the personified car 2.
  • the general-purpose text creation unit 230 of the first embodiment and the second embodiment is provided by the messaging platform of the messaging server 12c in the text input / output device 1c. That is, the general-purpose text generator 230 is provided as a common bot of the messaging platform. For this reason, in the present embodiment, the general-purpose text creation unit 230 is stored in the messaging platform.
  • the text selection unit 240 may be provided in the vehicular conversation server 10c.
  • the configuration of an information processing system Xd according to a fourth embodiment of the present disclosure will be described with reference to FIG. Also in this figure, the same reference numerals are added to the same components as in FIG. 2, FIG. 8, and FIG. 9, and the configurations of the control unit and the storage unit are omitted.
  • This information processing system Xd does not use a messaging platform. Therefore, the functions of the messaging according to the first to third embodiments are realized by the dedicated application of the terminal 3d. That is, in addition to the text acquisition unit 210 and the text output unit 250, the terminal 3d also includes the functions of the text reception unit 260 and the text transmission unit 270. Therefore, the information processing system Xd constitutes a text input / output device 1d including the vehicular conversation server 10d, the general conversation server 11d, and the terminal 3d.
  • the vehicle conversation server 10d and the general conversation server 11d may be the same as in the first embodiment.
  • the terminal 3 e includes the respective modules of the AI agent for realizing the vehicle text creating unit 220 and the general-purpose text creating unit 230, and the functions corresponding to the text acquiring unit 210, the text outputting unit 250, the text receiving unit 260, and the text transmitting unit 270 And have.
  • the state acquisition setting unit 200 is separately provided in the caching server e. That is, the information processing system Xe may constitute the text input / output device 1e by the terminal 3e and the caching server e. This is because the caching server e stores the vehicle data 300 as a cache because the terminal 3 e is not always connectable to the car 2.
  • the state acquisition setting unit 200 may be provided in the terminal 3 e so as to periodically acquire the vehicle data 300 without using the cache server 14 e.
  • Other Embodiments In the first to fifth embodiments of the present disclosure, the features of the present disclosure have been mainly described. However, the combination of the configurations of the functional units in these embodiments is arbitrary. That is, the configuration of any of the first to fifth embodiments, the configuration of any combination, or the configuration including all the functional units may be employed.
  • the text input / output device 1 and the terminal 3 of the present disclosure may further include functional units not described in the above first to fifth embodiments.
  • the information processing system of the present disclosure can be applied not only to a messaging platform serving as a remote operation interface of a vehicle, but also to an application of another portable terminal 3, an application of a PC (Personal Computer), a web application of WWW, etc. It is.
  • each section is expressed as S101, for example.
  • each section can be divided into multiple subsections, while multiple sections can be combined into one section.
  • each section configured in this way can be referred to as a device, a module, or a means.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Databases & Information Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Acoustics & Sound (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Mechanical Engineering (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • User Interface Of Digital Computer (AREA)
  • Machine Translation (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
PCT/JP2018/029218 2017-09-27 2018-08-03 情報処理装置、情報処理システム、及び情報処理方法 WO2019064928A1 (ja)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2017186429A JP6819529B2 (ja) 2017-09-27 2017-09-27 情報処理装置、情報処理システム、及び情報処理方法
JP2017-186429 2017-09-27

Publications (1)

Publication Number Publication Date
WO2019064928A1 true WO2019064928A1 (ja) 2019-04-04

Family

ID=65903067

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2018/029218 WO2019064928A1 (ja) 2017-09-27 2018-08-03 情報処理装置、情報処理システム、及び情報処理方法

Country Status (2)

Country Link
JP (1) JP6819529B2 (enrdf_load_stackoverflow)
WO (1) WO2019064928A1 (enrdf_load_stackoverflow)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112667091A (zh) * 2019-10-16 2021-04-16 北京搜狗科技发展有限公司 一种云端联想方法和相关装置
WO2025089205A1 (ja) * 2023-10-23 2025-05-01 矢崎総業株式会社 車両用対話システム
WO2025089204A1 (ja) * 2023-10-23 2025-05-01 矢崎総業株式会社 車両用対話システム

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP7351233B2 (ja) * 2020-02-13 2023-09-27 トヨタ自動車株式会社 プログラム、制御装置、及び制御方法
JP2025000512A (ja) * 2023-06-19 2025-01-07 ソフトバンクグループ株式会社 制御システム

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007511414A (ja) * 2003-11-20 2007-05-10 ボルボ テクノロジー コーポレイション 車両運転者および複数のアプリケーション間の対話方法およびシステム
JP2008026621A (ja) * 2006-07-21 2008-02-07 Fujitsu Ltd 音声対話機能を有する情報処理装置
JP2016212541A (ja) * 2015-05-01 2016-12-15 株式会社ＡＴＲ−Ｔｒｅｋ 対話装置、対話方法、及びそのコンピュータプログラム
JP2017067849A (ja) * 2015-09-28 2017-04-06 株式会社デンソー 対話装置及び対話方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6202041B2 (ja) * 2015-04-24 2017-09-27 マツダ株式会社 車両用音声対話システム

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007511414A (ja) * 2003-11-20 2007-05-10 ボルボ テクノロジー コーポレイション 車両運転者および複数のアプリケーション間の対話方法およびシステム
JP2008026621A (ja) * 2006-07-21 2008-02-07 Fujitsu Ltd 音声対話機能を有する情報処理装置
JP2016212541A (ja) * 2015-05-01 2016-12-15 株式会社ＡＴＲ−Ｔｒｅｋ 対話装置、対話方法、及びそのコンピュータプログラム
JP2017067849A (ja) * 2015-09-28 2017-04-06 株式会社デンソー 対話装置及び対話方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
MIYAZAWA: "Factors of Interaction in the Spoken Dialogue System with High Desire of Sustainability", THE IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS, COMMUNICATIONS AND COMPUTER SCIENCES ( JAPANESE, 1 January 2012 (2012-01-01), pages 27 - 36, ISSN: 0913-5707 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112667091A (zh) * 2019-10-16 2021-04-16 北京搜狗科技发展有限公司 一种云端联想方法和相关装置
WO2025089205A1 (ja) * 2023-10-23 2025-05-01 矢崎総業株式会社 車両用対話システム
WO2025089204A1 (ja) * 2023-10-23 2025-05-01 矢崎総業株式会社 車両用対話システム
JP2025071571A (ja) * 2023-10-23 2025-05-08 矢崎総業株式会社 車両用対話システム

Also Published As

Publication number Publication date
JP6819529B2 (ja) 2021-01-27
JP2019061546A (ja) 2019-04-18

Similar Documents

Publication Publication Date Title
JP7691523B2 (ja) 自動アシスタント応答を生成する際に大規模言語モデルを使用すること
US10839583B2 (en) Emotive advisory system and method
WO2019064928A1 (ja) 情報処理装置、情報処理システム、及び情報処理方法
US12148421B2 (en) Using large language model(s) in generating automated assistant response(s
US9085303B2 (en) Vehicle personal assistant
US9798799B2 (en) Vehicle personal assistant that interprets spoken natural language input based upon vehicle context
KR102562227B1 (ko) 대화 시스템, 그를 가지는 차량 및 차량의 제어 방법
KR102635811B1 (ko) 사운드 데이터를 처리하는 시스템 및 시스템의 제어 방법
KR20190041569A (ko) 대화 시스템, 이를 포함하는 차량 및 대화 서비스 처리 방법
CN113674742A (zh) 人机交互方法、装置、设备以及存储介质
US20240402989A1 (en) Human-computer interaction method and apparatus, and terminal device
US20200319841A1 (en) Agent apparatus, agent apparatus control method, and storage medium
KR20200006738A (ko) 대화 시스템 및 대화 처리 방법
CN117216334A (zh) 信息处理方法、装置、设备及存储介质
KR20220059629A (ko) 차량 및 그의 제어방법
US20250006202A1 (en) Biasing interpretations of spoken utterance(s) that are received in a vehicular environment
KR102487669B1 (ko) 대화 시스템, 이를 포함하는 차량 및 대화 처리 방법
KR20190135676A (ko) 대화 시스템, 이를 포함하는 차량 및 대화 처리 방법
JP2025128206A (ja) 自動アシスタント応答を生成する際に大規模言語モデルを使用すること
CN118587336A (zh) 一种交互式语音助手动画生成方法、系统、设备及介质

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18860707

Country of ref document: EP

Kind code of ref document: A1