JP4206818B2 - In-vehicle device and data creation device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an in-vehicle device and a data creation device, for example, an in-vehicle device provided with an agent function for performing conversation and communicating with a vehicle occupant autonomously and performing device operations, and the like. The present invention relates to a data creation device that creates a screen element transition body.
[0002]
[Prior art]
For example, pet-type robots such as dogs, and agent devices that interact with passengers by providing guidance on device operations such as navigation devices in the passenger compartment and asking questions and suggestions according to the situation have been developed. It is mounted on a vehicle as an in-vehicle device (see, for example, Patent Document 1).
In addition, various states that can be considered for the apparatus are assumed, and actions to be taken when each assumed state is detected are defined in accordance with predetermined data and programs.
[0003]
[Patent Document 1]
JP-A-11-37766
[0004]
In this conventional agent device, for example, the agent E appears on the display device 27 when the detection value G1 of the fuel detection sensor 415 is equal to or less than the average value G2 (G1 ≦ G2) of the remaining fuel amount for all five times. Then, a moving image for prompting refueling is displayed on the display device 27, and a sound such as “I am hungry! I want gasoline!” Is output from the sound output device 25.
[0005]
[Problems to be solved by the invention]
In the conventional agent device, the state in which the agent performs various operations corresponding to various situations is displayed as a moving image or a still image. For example, a series of videos that the agent bows, a nod video, a video that points to the answer by raising the right hand to prompt the user to select the answer to the question displayed on the screen, etc. It represents the operation of.
However, if the playback time of the moving image to be displayed is fixed, the following problem occurs. For example, in the case of an animation that prompts an operation, if the user hesitates to perform an operation or if the playback of a video ends while thinking about the operation, the user misunderstands that the operation cannot be continued. To give to. In this case, the user will not know what to do after that.
Also, when playing the same video for multiple situations, if the playback time of the video is fixed, the different audio contents output depending on the situation will be different, so the audio and video will not be synchronized It will also be.
[0006]
Therefore, it is conceivable to change the playback time of the moving image according to the situation.
For example, it is possible to adjust the time so that the entire moving image can be looped. However, if a situation occurs in which a transition to a subsequent video occurs during a loop, and the video is interrupted, if the position and orientation of the character at the time of the loop interruption differ from the first frame of the subsequent video, the subsequent video An operation gap occurs between them and the reproduction quality of the moving image is lowered. Also, depending on the operation content of the agent, for example, an unnatural repetition of the same operation, such as repeated bowing, may occur, giving the user a sense of discomfort.
On the other hand, it is conceivable to freeze the image without looping the movie, but the agent's movement is completely stopped, which may give the user a misunderstanding that the system has failed or stopped. .
Furthermore, it may be possible to prepare videos with different lengths depending on the playback time, but since it is necessary to have videos for each playback time, the amount of data to be stored increases and the time that has occurred since the start of video playback is uncertain It is impossible to make a moving image correspond to various situations (such as vehicle running state or communication state).
[0007]
In addition, when the user makes a selection while playing a moving image that requires the user to select, two methods can be considered as the timing for playing the subsequent moving image. First, in the case of a method in which a moving image being reproduced is immediately interrupted to shift to subsequent moving image reproduction, the operation becomes unnatural due to a gap between the interrupted image and the starting image of the subsequent moving image, and the reproduction quality is deteriorated.
On the other hand, in the case of a method in which a moving image being reproduced is reproduced to the end and a subsequent moving image is reproduced, the response of the agent's operation to the user's operation becomes worse.
[0008]
Therefore, in the present invention, when there is a predetermined instruction from the in-vehicle device or the user during execution of one screen element of the screen element transition body, unnaturalness with respect to the playback quality and response speed of the moving image stage being executed. It is a first object to provide an in-vehicle device capable of reproducing a subsequent moving image with a reduced amount.
In addition, the present invention creates a screen element transition body in which characters communicate with each other by being executed by the in-vehicle device, and a start condition thereof, and a predetermined instruction is given from the user or the in-vehicle device while executing the screen element. The second object of the present invention is to provide a data creation device capable of setting whether or not to interrupt the action expression of a character.
[0009]
[Means for Solving the Problems]
According to the first aspect of the present invention, the start moving image from the basic posture state of the character to the predetermined posture state indicating the expression content, the retained moving image holding the predetermined posture state, and the end moving image from the predetermined posture state to the basic posture state At least 3 videos Consist of Character motion expression Video Expression to memorize Video Storage means and action expression of the character Video Screen element transition that stores a screen element transition body composed of a combination of the screen elements, with one screen element defined as the display element containing the display content body Storage means, screen element transition body execution means for executing the screen element transition body, and one screen element of the screen element transition body Action expression video defined in If there is a predetermined instruction from the in-vehicle device or user during the execution of Concerned Running Action expression Ready mode to interrupt the video and execute the next screen element, Concerned Expression of action during execution Video And a mode determination means for determining a quality mode for executing the next screen element after the operation is completed.
According to a second aspect of the present invention, the mode determination means is a representation of an action being executed when a predetermined instruction is received from an in-vehicle device or a user. Video of 1 The in-vehicle device according to claim 1, wherein it is determined whether the mode is a quick response mode or a quality mode from a mode defined by a screen element.
In the invention according to claim 3, the mode determination means, when there is a predetermined instruction from the in-vehicle device or the user, In-vehicle The vehicle-mounted device according to claim 1, wherein whether the mode is an immediate mode or a quality mode is determined according to a table defined in the device.
In invention of Claim 4, A data creation device for creating a screen element transition body to be executed by an in-vehicle device, wherein a start moving image from a basic posture state to a predetermined posture state indicating an expression content, a holding moving image that holds the predetermined posture state, and the predetermined posture state Stipulated in a stage consisting of at least three videos from the end video to the basic posture state Character motion expression Video A screen element with specified display content including As one screen element From the screen element creation means to be created and the one screen element created by the screen element creation means, the following 1 A transition condition setting means for setting a transition condition for shifting to a screen element; 1 Screen element transition body creating means for creating a screen element transition body to be executed by the in-vehicle device based on the screen element and the transition condition, the screen element creation means, 1 Screen element Action expression video defined in In-vehicle device Or its user Character's motion expression when given instructions from Video Whether to interrupt About mode The , In the one screen element Provided is a data creation device further comprising operation interruption setting means for setting.
According to a fifth aspect of the present invention, the operation interruption setting means is an operation being executed. Expression video The Interruption Responsive mode to execute the next screen element, and motion expression during execution Video Quality mode, which executes the next screen element after continuing Whether to interrupt In-vehicle device side Judging by Automatic judgment mode Any one of the above is set, and the data creation device according to claim 4 is provided.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 to FIG. 1 show an agent device that is a preferred embodiment of the in-vehicle device of the present invention, a scenario creation device that is a preferred embodiment of a data creation device, and a scenario editor that is a preferred embodiment of a data creation program. This will be described in detail with reference to FIG.
[0011]
(1) Outline of the embodiment
In the agent device of the present embodiment, an agent (character) image (planar image, stereoscopic image such as holography) having a predetermined appearance is displayed in the vehicle. The function of the agent device is a function of recognizing and judging the surrounding state (including human movement and voice) from the detection result of the sensor or the like, and outputting a video or voice of the action according to the result. It is executed in conjunction with the movement and voice of this agent's appearance. For example, an operation of tilting the head is performed together with a question requesting an answer (Japanese food, Western food, etc.) such as “Which genre do you like?”. Then, the content of the user's answer to this question is discriminated (recognized from the answer voice recognition or the answer selection button 54a selection), and the process corresponding to the next scene (screen element) is executed. As described above, since the inquiry requesting an answer from the apparatus and the execution of the predetermined operation is started in response to the answer, the user seems to have an agent with a pseudo personality in the vehicle. Come to experience. In the following description, execution of a series of functions of such an agent device will be described as an agent action or operation.
[0012]
In the agent device of the present embodiment, this agent is allowed to perform various communications with the driver and proxy operations. Various actions (each action) autonomously performed by the agent are configured by a plurality of scenarios (screen element transition bodies). And, autonomously start (activate) a plurality of scenarios that define the contents of a series of continuous actions (reproduction of moving images representing operations and output of sound, and operations with agents), and the deployment of each scenario. The scenario data standardized by the autonomous activation condition (activation condition) for saving is saved.
change.
[0013]
A scenario is created by a scenario creation device and is composed of one or a plurality of continuous scenes with a scene (screen element) as a reference unit. One scene is composed of at least one of a processing content autonomously processed and a moving image and an audio representing an agent's operation. Which scene to transition to after each scene is determined by the transition conditions.
[0014]
The action of the character in each scene is composed of one or a plurality of individual action images (character action expressions), and each individual action image is a moving image of a unit capable of expressing at least one complete meaning. This individual motion image is divided into a starting moving image from the character's basic posture state to a predetermined posture holding state (predetermined state), a holding moving image that performs a slight motion within a predetermined range from the holding state, and the holding state to the basic posture state. It consists of three of the end videos until.
Thus, the continuity of motion can be maintained by configuring each individual motion image of the character with a moving image that returns from the basic posture state to the basic state through a predetermined state.
Then, the reproduction time of the individual motion image is adjusted by repeating the retained moving image. In this case, the retained moving image returns to a predetermined state after performing an operation such as shaking hands or blinking, so that the user can recognize that the agent is operating. Further, when the retained moving image has a slight movement in the retained state, the gap between images does not increase even if the retained moving image is interrupted in the middle of repetitive reproduction and is shifted to the end moving image.
When the response speed is more important than quality, the individual operation image being reproduced is interrupted when the interruption condition (the predetermined instruction from the in-vehicle device or the user (see FIG. 13)) is satisfied, and the next operation image is interrupted. By reproducing the individual motion image, the response of the character can be made quick.
[0015]
The agent device determines how to reproduce the subsequent individual motion image when the interruption condition is satisfied during the reproduction of the individual motion image, the quick response mode defined in the scene (screen element), the quality mode, Determine according to the automatic selection mode. When the automatic selection mode is selected, the quick response mode or the quality mode is determined in advance from a table corresponding to each interruption condition.
[0016]
The user of the agent device creates a unique scenario using the scenario creation device in accordance with a defined standard. The scenario creation device can be configured by installing a scenario editing program and data in a personal computer.
The created scenario is transmitted to the agent device via a network such as the Internet or downloaded, and stored in the agent device via a predetermined semiconductor memory. It is possible to make the agent perform the desired action (communication and processing). Furthermore, the created scenario can be attached to an e-mail and transmitted to the agent device.
In this way, the user can easily and uniquely create a scenario that causes the agent to function as intended, and thus there is no resistance to the autonomous operation of the agent device.
In the scenario creation device, when creating a scenario, it is possible to specify the interruption condition for each individual motion image in each scene and the mode for shifting to the subsequent individual motion image after the interruption in each scene data. ing.
[0017]
(2) Details of the embodiment
FIG. 1 shows an overall system configuration including an agent device and a scenario creation device.
In this system, the agent device 1 of the present embodiment, a scenario creation device 2 of a scenario data creator who is a user or a third party who creates scenario data according to a specified standard, and the Internet using a server 3 or the like It consists of communication means.
The scenario creation device 2 creates unique scenario data using a scenario editor. The user who created the unique scenario data can store the scenario data in a semiconductor storage device or other storage medium 7 such as a DVD-ROM or an IC card and transfer it to the agent device 1. The agent device 1 that has received the scenario data reads the scenario data from the storage medium 7 by the storage medium driving device and incorporates the scenario data into the already stored scenario data, whereby the scenario data created by the scenario creation device 2 is converted into the scenario data. Therefore, the agent device 1 can be operated. Note that what is created by the scenario creation device 2 may be the user of the agent device 1 or a third party.
Further, the agent device 1 can incorporate scenario data created by the user or a third party via a network such as the Internet, or can incorporate scenario data attached to an email.
In addition, a third party who desires to provide a service to the user of the agent device 1 creates scenario data in a predetermined format by the scenario creation device 2 using, for example, a scenario editor and posts it on the homepage. It can be downloaded or sent to the agent device 1 as an email attachment. The agent device 1 receives the scenario data 5 attached to the e-mail, or the user downloads the scenario data file 4 via communication means such as the server 3. Further, the agent device 1 transmits the user's response (response email to the scenario data) acquired in accordance with the execution of the received scenario data to the scenario creator 2 of the scenario creator in the body of the email 6 or an attached file. .
[0018]
First, the configuration and operation of an agent device 1 in which an agent functions autonomously according to a scenario created by a developer or a user will be described.
FIG. 2 is a block diagram showing the configuration of the agent device 1 in the present embodiment.
The agent device 1 according to this embodiment is mounted on a vehicle and travels to a user in addition to an agent function such as a function of communicating with a user in the vehicle and a vehicle control function of performing predetermined processing on the vehicle. It also has a navigation function for route guidance.
In the agent device 1 of the present embodiment, a central processing device (1), a display device (2), a voice output device (3), a voice input device (4), an input device (for realizing an agent function and a navigation function) 5), various status detection devices (6), various in-vehicle devices (7), a communication control device (8), a communication device (9), and an external storage device (10).
[0019]
The central processing unit (1) includes a CPU (1-1) that executes various arithmetic processes, a flash memory (1-2) that reads and stores a program from the external storage device (10), and a flash memory (1-2). ROM (1-3) storing a program (program reading means) for performing a program check and update process, and a RAM (1-4) for temporarily storing data being processed by the CPU (1-1) as a working memory ), Clock (1-5) used to measure the passage of time to reduce the element value of short-term emotion elements over time, and other time and time measurements, and display on the display device (2) Image data (1-7) in which image data to be used is stored, and image data is extracted from the image memory (1-7) based on the display output control signal from the CPU (1-1), and subjected to image processing. From the image processor (1-6) and CPU (1-1) for output to the display device (2) An audio processor (converting an audio output control signal into an analog signal and outputting the analog signal to the audio output device (3)) and an analog processor input from the audio input device (4) into a digital audio input signal ( 1-8), input device I / F unit (1-9) for receiving the input content from the input device (5), various input I / F units for receiving information from detectors for detecting various situations ( 1-10) Reading data and programs from an external storage medium (10-2) such as a communication I / F section (1-11) for exchanging information with other devices, CD-ROMs, IC cards, hard disks, etc. An external storage device control unit (1-12) is provided for controlling the external storage device (10) for writing or writing data.
[0020]
This central processing unit (1) includes route search processing, display guidance processing necessary for route guidance, other processing necessary for the entire system, agent processing in this embodiment (various communication and operation agency between agents and drivers) The situation determination is performed, and processing autonomously performed according to the result is performed.
The program (program reading means) for performing the update process may be stored in the flash memory (1-2) in addition to the ROM (1-3).
All the programs executed by the CPU (1-1) including the program in the present embodiment may be stored in a CD-ROM or the like that is an external storage medium (10-2), or a part of the programs Alternatively, all may be stored in the ROM (1-3) or the flash memory (1-2) on the main body side.
The data and programs stored in the external storage medium (10-2) are input as external signals to the central processing unit (1) and processed to realize various agent functions and navigation functions. It has become.
Further, the central processing unit (1) of the present embodiment forms a screen element transition body execution means for executing a screen element transition body (scenario) when it is determined that the start condition (autonomous start condition) is satisfied. is doing.
[0021]
The display device (2) displays a road map for route guidance and various image information by the processing of the central processing device (1), and screen element transitions composed of various character actions (videos) and screen configuration parts. The body (scenario) is displayed. Various display devices such as a liquid crystal display device and a CRT are used for the display device (2). The display device (2) may have a function as an input device (5) such as a touch panel.
The voice output device (3) provides guidance voice when performing route guidance by voice by the processing of the central processing unit (1), a conversation for normal communication with the driver by the agent, and questions for obtaining driver information. Voice and sound are output. The audio output device (3) is composed of a plurality of speakers arranged in the vehicle. These may also be used as audio speakers.
[0022]
In the voice input device (4), a dedicated microphone having directivity is used in order to accurately collect the voice of the driver. A voice recognition process is executed by the CPU (1-1) using a digital voice input signal obtained by converting an analog signal inputted from the voice input device (4).
Examples of voices that are subject to voice recognition include input voices such as destinations in navigation processing, driver conversations with agents (including responses by drivers), and the voice input device uses these voices. Functions as voice input means for inputting.
As to whether or not a scene requires voice recognition, a voice recognition instruction is set in each scene data. In addition, a dictionary for recognizing a voice that is a target of voice recognition is specified in the scene data of a scene for which a voice recognition instruction is set.
In some scenarios, an instruction to change the element value of the emotional element of the agent may be defined in accordance with the result of the voice recognition (the driver's response result).
[0023]
The input device (5) is used to input a telephone number or coordinates on a map when setting a destination, or to request (request) a route search or route guidance to the destination. The input device (5) is used as a trigger when the driver inputs driver information or when the use of the agent function is started. Furthermore, the input device (5) functions as one response means for the driver to respond to an inquiry from the agent or the like in communication with the agent by the agent function.
As the input device (5), various devices such as a touch panel (functioning as a switch), a keyboard, a mouse, a light pen, and a joystick can be used.
Moreover, you may provide the receiving part which receives the remote control using infrared rays etc., and the various signals transmitted from a remote control.
Further, voice recognition using the voice input device (4) may be used instead of the input device.
[0024]
FIG. 3 is a block diagram showing the configuration of various status detection devices (6).
The various situation detection devices constitute situation detection means for detecting various on-vehicle situations.
Various situation detection devices (6) are a current position detection device (6-1), a traffic situation information reception device (6-2), and a brake detector (6-3) Side brake (parking brake) detector (6-4), accelerator opening detector (6-5), A / T shift position detector (6-6), wiper detector (6- 7), a direction indicator detector (6-8), a hazard detector (6-9), and an ignition detector (6-10). With the above configuration, detection means is formed by detecting various situations and conditions.
Each situation detection device (6) includes a vehicle speed sensor (6-11) for detecting the speed of the vehicle (vehicle speed information), and whether the vehicle is traveling depending on whether or not the vehicle speed detected by the vehicle speed sensor is zero. By determining whether or not, a travel determination means is formed.
[0025]
The current position detection device (6-1) is for detecting the absolute position (by latitude and longitude) of the vehicle, and is a GPS (Global Positioning System) reception device that measures the position of the vehicle using an artificial satellite. (6-1-1), a data transmission / reception device (6-1-2) that receives a GPS correction signal, a direction sensor (6-1-3), and a steering angle sensor (6-1-4) A distance sensor (6-1-5) is used.
The distance sensor (6-1-5) and the rudder angle sensor (6-1-4) also function as driving operation status detection means.
[0026]
The traffic condition information receiving device (6-2) is for detecting the congestion condition of the road.
The traffic condition information receiving device (6-2) includes a beacon receiving device (6-2-1) that receives information from beacons placed on the road, and a device (6-2) that receives information using FM broadcast radio waves. -2) etc. are used to receive traffic jam information and traffic regulation information from the traffic information center.
Further, the beacon receiving device (6-2-1) may be used in combination with the current position detecting device (6-1) as the current position detecting means.
[0027]
The brake detector (6-3) detects whether or not the foot brake is depressed.
The side brake (parking brake) detector (6-4) detects whether or not the driver is operating the side brake and the state of the side brake (ON or OFF).
The accelerator opening detector (6-5) detects how much the driver has depressed the accelerator pedal.
The shift position detector (6-6) detects whether or not the driver is operating the A / T shift lever and the shift lever position.
The wiper detector (6-7) detects whether or not the driver is using the wiper.
[0028]
The direction indicator detector (6-8) detects whether or not the driver is operating the direction indicator and whether or not the direction indicator is blinking.
The hazard detector (6-9) detects whether or not the driver is using the hazard.
The ignition detector (6-10) detects whether or not the ignition switch is ON.
A distance sensor (6-1-5) can also be used to detect the vehicle speed.
In addition to these, the various state detection device (6) detects a light detection sensor for detecting an operation state of lamps such as a headlamp and a room lamp, and a driver's seat belt attaching / detaching operation. A seat belt detection sensor and other sensors are provided.
[0029]
The GPS receiver (6-1-1), the data transmitter / receiver (6-1-2), and the traffic information receiver (6-2) are connected to the communication device I / F unit (1-11) in FIG. Others are connected to various input I / F units (1-10).
[0030]
In FIG. 2, a communication control device (8) can be connected to the communication device I / F unit (1-11). The communication control device (8) is connected to a communication device (9) (such as a mobile phone made up of various wireless communication devices).
Using these, in addition to telephone line calls, for example, communication with information providing stations that provide karaoke data used for communication karaoke in the car, information base stations that provide traffic information, and agent processing It is also possible to enable communication with an information provider that provides scenario data to be used.
[0031]
In the present embodiment, the central processing unit (1) can receive an e-mail attached with a scenario via the communication control unit (8).
In addition, the central processing unit (1) can incorporate browser software for displaying a homepage on the Internet and can be processed by the CPU (1-1). The scenario can be executed from the homepage via the communication control unit (8). You can download the data including.
The communication control device (8) may be integrated with the communication device (9).
[0032]
The central processing unit (1) receives in-vehicle communication through the communication I / F unit (1-11) to receive the operation status of the other in-vehicle device (7) and performs various controls on the in-vehicle device. It is like that.
For example, the central processing unit (1) controls the air conditioner so as to raise or lower the set temperature of the air conditioner that is the various in-vehicle devices (7). In addition, the driver controls the audio device such that the driver increases or decreases the output volume of an audio device such as a radio, a CD player, or a cassette player from the audio device. The control for these in-vehicle devices is performed along with the execution of the scenario when the control for the in-vehicle devices is defined in the scenario.
[0033]
The external storage device (10) includes an external storage medium driving unit (10-1) and its external storage medium (10-2). The external storage device (10) reads data and programs from the external storage medium (10-2) under the control of the external storage device control unit (1-12) according to instructions from the CPU (1-1), and Data and programs are written to the external storage medium (10-2).
As the external storage medium (10-2), for example, various storage media such as a flexible disk, a hard disk, a CD-ROM, a DVD-ROM, an optical disk, a magnetic tape, an IC card, and an optical card are used. Each external storage medium driving device (10-1) is used.
[0034]
A plurality of external storage devices (10) may be possessed in the system. For example, the collected personal information such as driver information data (10-2-3-6), learning item data and response data (10-2-3-7) can be easily carried with an IC card or flexible disk. An example in which the data is configured and other data is configured by a DVD-ROM is conceivable. By doing this, when driving other vehicles, it is possible to read and use the data from the IC card that stores them, and to communicate with the agent who has learned the situation that the user has responded to in the past Become. That is, it is possible to make an agent having learning content unique to each driver appear in the vehicle instead of the agent for each vehicle.
Further, even when the configuration includes the scenario data + image data (10-2-3-4) used in the scenario as an example in a DVD-ROM, it can be added using an IC card.
Thereby, it is possible to add an original scenario unique to each user.
Thus, by storing the screen element transition body (scenario) and the activation conditions of the screen element transition body from the outside, the screen element transition storage means of the present invention is formed, the screen configuration including the character image, and the character By storing the control contents to be executed including the display, the storage means of the present invention is formed.
[0035]
CPU (1-1) is a program (10-2-1) that realizes various agent functions and navigation functions, and agent data (10-2-3) and navigation data (10-2-2) used for arithmetic processing Is stored (installed) in a separate external storage device (for example, a hard disk device) from the DVD-ROM or IC card shown in the above configuration example, and necessary programs are stored in the flash memory (1-2). The data may be read (loaded) and executed, or data necessary for arithmetic processing may be read (loaded) from the storage device into the RAM (1-4) and executed.
[0036]
Next, the configuration of a program executed by the CPU (1-1) including the program according to the present invention will be described.
FIG. 4 shows the relationship between the agent processing unit (101) realized by the CPU (1-1) executing a program and the overall processing unit (102).
In this embodiment, an agent function navigation device is realized by adding an agent processing unit (101) for realizing an agent function to an overall processing unit (102) for realizing various navigation functions.
[0037]
Each of the agent processing unit (101) and the overall processing unit (102) has an I / F unit for exchanging the processing data with each other, and can acquire the processing data of each other.
For example, when the agent processing unit (101) acquires destination data that the driver wants to set as a result of executing communication with the driver according to the scenario data, the agent processing unit (101) supplies this data to the overall processing unit (102). It has become.
The overall processing unit (102) searches for a route based on the acquired destination data and performs route guidance based on the generated travel route data. In this route guidance process, when guidance such as a route change direction by image or voice is provided, the data required for guidance is supplied from the overall processing unit (102) to the agent processing unit (101) to guide the traveling route It is also possible for the agent to guide according to the scenario data converted into data.
[0038]
FIG. 5 shows the configuration of the agent processing unit (101).
The agent processing unit (101) includes a scenario driving unit (101-1), an autonomous activation determination unit (101-2), a learning unit (101-3), a character psychology unit (101-4), a drawing / A voice output unit (101-5), a voice recognition unit (101-7), an agent OS unit (101-8), and an external I / F unit (101-9) are provided.
The scenario driver (101-1) reads the scenario data (10-2-3-4) and instructs each processing unit using message communication based on the scenario data (functions provided by each processing unit). use. The scenario driving unit (101-1) performs central processing of the agent processing unit such as managing scenario execution and providing various agent functions to the driver.
[0039]
The autonomous activation determination unit (101-2) holds autonomous activation condition data for each scenario in the scenario data (10-2-3-4), and is periodically output from the agent OS unit (101-8). Various conditions and conditions such as time, location where the vehicle is located, road type such as general road and highway, vehicle state such as running or stopped, operating state such as operating navigation device or guiding, etc. Comparison and judgment.
When the conditions are matched, the autonomous activation determining unit (101-2) issues an instruction to execute the scenario having the matched conditions to the scenario driving unit (101-1).
Various situations for comparison with the autonomous activation conditions are obtained from the agent OS unit (101-8) and the learning unit (101-3).
[0040]
The learning unit (101-3) in FIG. 5 uses the driver information data (10-2-3-6), the items (execution results and execution history) obtained by the driver's selection and response in communication with the agent, Store as learning item data and response data (10-2-3-7). The learning unit (101-3) also obtains an end ID indicating how to end when the scenario ends in a different scene, and stores it as response data (10-2-3-7). These acquired items are stored on the RAM (1-4), but can be output to an IC card or the like as an external storage medium (10-2).
The learning unit (101-3) obtains a change in the situation from the agent OS unit (101-8) and records information related to the driving operation. For example, the date and time of power ON (ignition ON) may be stored for the past 10 times in order to determine various conditions such as the boarding time zone and boarding frequency by the driver. The stored information is provided to the scenario driver (101-1), for example, and used to change the development of the scenario, or used for comparison of autonomous activation determination.
In addition, although the learning part (101-3) in this embodiment serves also as holding | maintenance and reference of driver information, you may make it independent as a driver information part.
[0041]
The character psychology unit (101-4) obtains the current situation managed by the agent OS unit (101-8), and sets long-term emotion change conditions and short-term emotion change conditions (FIGS. 8 and 10) described later. Based on the character's psychological state, the long-term emotional element and the short-term emotional element are autonomously changed.
Also, the character psychology unit (101-4) obtains a mental model change instruction (instruction for changing the emotion element value of the character) in the scenario from the agent OS (101-8), and determines a long-term emotion element according to the change instruction. , Change short-term emotional elements.
[0042]
The drawing / sound output unit (101-5) creates a control signal for displaying a screen composed of parts such as a selection button and a title in response to an instruction from the scenario driving unit (101-1). In addition, a control signal for displaying various actions (motions) of the character corresponding to the display state by the scene data is also created by an instruction from the scenario drive unit (101-1).
In this embodiment, these control signals are transmitted to the agent OS unit (101-8), transmitted from the external I / F unit (101-9) to the overall processing unit (102), and are present in the overall processing unit (102). The image processor (1-6) is transmitted to the image processor (1-6) through the processing unit for instructing the image processor, and the image is processed and displayed on the display device (2). In 8), a processing unit for giving an instruction to the image processor may be provided.
[0043]
The drawing / voice output unit (101-5) also creates a control signal for outputting a dialogue when the agent communicates with the driver in response to an instruction from the scenario driving unit (101-1).
In this embodiment, these are transmitted to the agent OS unit (101-8), transmitted from the external I / F unit (101-9) to the overall processing unit (102), and sent to the voice processor in the overall processing unit (102). The signal is transmitted to the audio processor (1-8) through the processing unit for giving instructions, and this audio output control signal is converted into an analog signal and output to the audio output device (3), but without passing through the overall processing unit (102). The agent OS unit (101-8) may have a processing unit that gives instructions to the voice processor.
The drawing / sound output unit (101-5) of the present embodiment includes a character motion drawing function and a sound output function in each scene, but a drawing unit (drawing function unit) and a sound output unit (sound The output function unit) may be configured separately.
[0044]
The voice recognition unit (101-7) issues a control signal for causing the voice recognition processing unit in the overall processing unit (102) to create a voice recognition dictionary in response to an instruction from the scenario driving unit (101-1). The voice recognition unit (101-7) also issues a control signal for starting and stopping the voice recognition process in response to an instruction from the scenario driving unit (101-1).
In the present embodiment, these are transmitted to the agent OS unit (101-8) and from the external I / F unit (101-9) to the speech recognition processing unit in the overall processing unit (102).
The voice recognition processing unit transmits an instruction to start and stop the voice recognition process to the voice processor (1-8), and the voice processor (1-8) receives an analog signal input from the voice input device (4). Is to be converted into a digital audio input signal.
When a voice input signal is input, the voice recognition processing unit acquires the digital voice input signal, and based on that, the voice recognition processing unit performs a recognition process, and the result is the reverse of the previous route. Is transmitted to the voice recognition unit (101-7). The speech recognition unit (101-7) notifies the scenario recognition unit (101-1) of the speech recognition result.
With the above configuration, voice recognition means for recognizing voice is formed.
[0045]
The agent OS unit (101-8) obtains changes in the situation (including the addition of scenarios) such as time, place, and various inputs, manages the current situation, and sends messages as necessary to the change in situation. It notifies each processing unit such as the character psychology unit (101-4) through communication. The change in the situation is supplied from the overall processing unit (102) through the external I / F unit (101-9) or obtained through an inquiry.
The information obtained is obtained by fetching the detection results by various status detection devices (6) from various input I / F units (1-10) and communication I / F units (1-11), and RAM (1-4). Is written on. Content input using the input device (5) is also supplied from the overall processing unit (102) through the external I / F unit (101-9), and the content is sent to each processing unit by message communication as necessary. Notice.
The agent OS unit (101-8) also has various other libraries, providing message communication for exchanging data between processing units, providing the current time, and managing memory. It provides memory necessary for each processing unit to perform processing, and provides data reading and writing functions from an external storage medium.
[0046]
Further, the agent OS unit (101-8) performs time processing using time information acquired from the clock (1-5) and serves as a timer to notify the passage of a specific time. That is, the agent OS unit (101-8) functions as a time measuring means, and measures the timer setting time set in each scene of the scenario. The timer setting time to start timing is notified from the scenario driving unit (101-1), and when the timer setting time has elapsed, the agent OS unit (101-8) indicates that the setting time has elapsed. -1) is notified.
In addition, the agent OS unit (101-8) starts timing in response to a short-term emotional element change notification from the character psychology unit (101-4), and character psychology every elapse of a predetermined time (for example, 3 minutes). Section (101-4) is notified of the time information. The character psychology section (101-4) decreases the value of the short-term emotion element by a predetermined value (for example, “3”) every time time information is notified, and when the element value becomes “0”, An end instruction is notified to the agent OS unit (101-8).
[0047]
The agent OS unit (101-8) periodically issues an autonomous activation determination instruction to the autonomous activation determination unit (101-2). This periodic autonomous activation determination instruction is issued every predetermined time. The predetermined time may be as short as possible within a range in which the autonomous activation determination process periodically processed by the autonomous activation determination instruction issued periodically does not affect other processes of the entire central processing unit (1). Desirably, the interval is set to 5 seconds in this embodiment. The predetermined time may be arbitrarily changed by the user by an operation from the input device (5).
Also, the agent OS unit (101-8) issues an autonomous activation determination instruction to the autonomous activation determination unit (101-2) even when it is determined that the change in the situation is large. Cases where the change in the situation is considered to be large include, for example, when the driver has set a destination, the vehicle has been removed from the guide route, scenario data has been added, scenario data has been deleted, etc. Yes, corresponding items are defined in advance and stored in the RAM (1-4) or the like.
[0048]
The external I / F unit (101-9) serves as an interface between the agent processing unit (101) and the overall processing unit (102). Part exists). Acquisition of various types of information such as navigation information used in agent processing, and transmission of control signals from the agent processing unit to the entire processing unit control navigation.
Drawing instructions to the image processor (1-6) and audio output to the audio processor (1-8), which are performed by notifying the overall processing unit (102) through the external I / F unit (101-9) The agent processing unit has a processing unit that gives instructions to other processors and the I / F unit, such as instructions and acquisition of input information from the input device I / F unit (1-9). You may make it.
[0049]
An overall processing unit (102) in FIG. 4 includes a map drawing unit, a route search unit, a route guidance unit, a current position calculation unit, a destination setting operation control unit, etc. (not shown), and an application unit that performs navigation signal output processing, In addition, it is configured by an OS unit or the like such as a program for performing display output control necessary for map display and route guidance and voice output control necessary for voice guidance.
The overall processing unit (102) also includes a speech recognition processing unit that performs speech recognition and a processing unit that converts text data into speech data. When the browser function or the mail function is added, the processing unit is added to the overall processing unit (102).
Alternatively, the agent processing unit (101) may have a browser function or a mail function.
In this embodiment, an extended function for executing agent processing is added to the overall processing unit (102). This extended function includes, for example, means for detecting the type of road that is running (highway, national road, etc.) from the road data in the navigation data and the current position, and the curve condition of the road that is running (before the curve, Means for detecting (end of curve) and the like exist.
These detected situations are transmitted to the agent processing unit, and are used, for example, to change an agent's emotional element.
[0050]
Next, the data configuration (including programs) stored in the external storage medium (10-2) will be described.
FIG. 6 conceptually shows information recorded in the external storage medium (10-2).
The external storage medium (10-2) includes a program (10-2-1) for realizing various agent functions and navigation functions according to the present embodiment, and agent data (10-2-3) and navigation as necessary various data. Data (10-2-2) is stored.
The navigation data (10-2-2) is composed of various data necessary for map drawing, route search, route guidance, destination setting operation, and the like. Examples include map data used for route guidance (road maps, housing maps, building shape maps, etc.), intersection data, node data, road data, photo data, registration point data, destination point data, guide road data, It consists of detailed destination data, destination reading data, telephone number data, address data, and other data files and stores all data necessary for the navigation device. Moreover, communication area data etc. are also memorize | stored as needed.
[0051]
Agent data (10-2-3) includes mental model data (10-2-3-1), recommended proposal data (10-2-3-3), and knowledge data (10-2-3-2). ), Scenario data (10-2-3-4), character data (10-2-3-5), driver information data (10-2-3-6), learning item data and response data (10-2-3-7) and operation switching judgment data (10-2-3-8).
[0052]
FIG. 7 conceptually shows the contents of the mental model data (10-2-3-1). The long-term emotional element 10-2-3-1a and the long-term emotional change condition 10-2-3 -1b, short-term emotional element 10-2-3-1c, and short-term emotional change condition 10-2-3-1d are stored.
The long-term emotion element 10-2-3-1a and the short-term emotion element 10-2-3-1c store element values representing the character's psychological state. The long-term emotional change condition 10-2-3-1b and the short-term emotional change condition 10-2-3-1d are the values of each emotional element depending on the detection values of the various situation detection devices (6) and the data by the navigation function. The condition for changing the value and the change value are stored.
The emotional element values of the long-term emotional element 10-2-3-1a and the long-term emotional change condition 10-2-3-1b are determined according to the change value corresponding to the change condition and the value of the emotional element change instruction according to the scenario. Is changed.
[0053]
FIG. 8 conceptually shows the long-term emotional element 10-2-3-1a.
As shown in FIG. 8, the long-term emotional element 10-2-3-1a is represented by each element of friendship, obedience, confidence, morality, and spirit, and each emotional element value is, for example, 0 It is represented by a value of ~ 100. The number of long-term emotional elements is five in this embodiment, but other predetermined numbers may be added by adding or omitting other elements.
The value of each element changes independently from 0 to 100.
When the long-term emotional element 10-2-3-1a satisfies the change conditions described in 10-2-3-1b, and when the short-term emotional change is instructed in the scenario being executed Each of them changes (updates) by a corresponding value.
[0054]
The long-term emotional element 10-2-3-1a has a reference value of 50 for each emotional element. When the reference value is higher than that, it indicates that the state is strong, and when it is lower, the state is weak.
Corresponding to the branch condition (transition condition) according to the scenario, as shown in FIG. 8B, each element value has five levels of “very low” “low” “normal” “high” “very high”. It is divided into In addition, it is good also as three steps (width | variety of each element value = 10) except "very low" and "very high". In addition, regarding the range of each divided element, the user of the agent device may be able to change the setting value. In this embodiment, the emotional elements are hardly changed (a large number of normal areas are set), but the distribution may be different from that of the present embodiment, such as equal allocation.
In the scenario creation device, five long-term emotional elements can be used for scene branching, and individual emotional element values may be used alone. Condition).
For example, if the condition when the agent greets and speaks vigorously is only when the degree of friendship is high, the humanity can be expressed more.
[0055]
Although not specifically shown, the long-term emotional change condition 10-2-3-1b includes the following descriptions of the navigation function, vehicle state, and other conditions and conditions due to changes (item column), and emotions that change in that case Element change values are specified.
For example, there is a rule such as “when the vehicle suddenly accelerates, the agent's emotion changes (friendship and morality are reduced by 1)”.
The change value of each long-term emotional element stipulated in the long-term emotional change condition 10-2-3-1b changes long-term emotions. It is desirable to change with the value.
[0056]
FIG. 9 is a conceptual representation of the short-term emotional element 10-2-3-1c.
As shown in FIG. 9, the short-term emotional elements are composed of four elements: joy, anger, sadness, and surprise. Although the number of short-term emotional elements is four in the present embodiment, other predetermined elements may be added or omitted by adding other elements.
Each value of the short-term emotional element 10-2-3-1c varies from 0 to 100, and only one element of each element takes a value greater than zero. That is, when a certain emotional element is newly changed from 0 to a predetermined value (for example, 50), other emotional elements having values before the change are changed to 0.
In this way, short-term emotional elements are the emotions that agents have in the short-term at that point in time for a certain situation or state, so previous values do not affect or accumulate like long-term emotional elements Only one element takes a value. In other words, humanity is expressed more by preventing multiple short-term emotions from being side by side.
However, other changes may be defined. For example, when element A is going to change 50, element A changes to 50 if other element B before the change is 50 or less, but element A changes if element B before the change is greater than 5. Do not. In this case, of the value of the element B before the change and the value of the element A to be changed, the element having a larger value may be changed to a value obtained by subtracting the smaller value. For example, if element A is 50 and element B is 80, element A remains 0 and element B decreases to 30, and if element A is 90 and element B is 40, element A changes to 50 B changes to 0.
[0057]
The short-term emotional element 10-2-3-1c satisfies the change conditions described in 10-2-3-1d, and the short-term emotional change is instructed in the scenario being executed Each of them changes (updates) by a corresponding value.
Also, the value of the newly changed short-term emotion element decreases with the passage of time and finally becomes zero. For example, if the maximum duration of emotion is defined as 1 hour, the time when the value decreases from the maximum value 100 to the minimum value 0 is 1 hour, so the value decreases by 5 every 3 minutes. Become. However, the setting may be changed so as to change by a predetermined value n in units of a predetermined time t.
In this way, the value of each emotional element in the short-term emotional element 10-2-3-1c corresponds to the human emotion that the emotional highness momentarily occurs and subsides over time, and the overall change range It changes with a large value up to a maximum of 100 percent, and gradually decreases by a predetermined value n at a predetermined time t interval. This makes the short-term change of the agent closer to humans.
[0058]
When each of the emotional elements of the short-term emotion element 10-2-3-1c has a value greater than 0 (for example, joy), the agent's short-term emotion becomes an element of that value (joy). If all short-term emotional elements are all zero, the agent's short-term emotion is “normal”.
The reference value of each emotional element in the short-term emotional element 10-2-3-1c is 0, and the higher the value is, the stronger the state is.
The short-term emotional element 10-2-3-1c is divided into three stages, “small”, “medium”, and “large”, as shown in FIG. It is divided. In addition, it is good also as five steps (width | variety of each element value = 20) which added "very small" and "very large". In addition, regarding the range of each divided element, the user of the agent device may be able to change the setting value.
[0059]
Although not shown in particular, the contents of the short-term emotion change condition 10-2-3-1d include the navigation function, the description of the condition such as the vehicle state and conditions (item column), and the emotion that changes in that case Element change values are specified.
For example, there is a rule such as “when the scenario is forcibly terminated, the agent's emotion changes (sadness is increased by +30)”.
[0060]
The change values of the short-term emotional elements stipulated in the short-term emotional change condition 10-2-3-1d do not take negative values unlike the long-term emotional change conditions, and are all positive values. Further, as described above, a large value (30, 70, 100 in this embodiment) up to a maximum of 100 percent is defined with respect to the entire change range (0 to 100 in this embodiment).
[0061]
In FIG. 6, the recommended proposal data (10-2-3-3) is used when a restaurant or the like is proposed as recommended information to the driver. This recommended proposal data (10-2-3-3) is composed of restaurant name, reading data, restaurant genre data, atmosphere data, fee data, point data, etc., and driver information data (10-2-3). Based on 3-6) and knowledge data (10-2-3-2), a restaurant recommended for the driver is searched and proposed. In addition to restaurants, there are sightseeing spots and rest areas.
Knowledge data (10-2-3-2) is based on statistical data, preference trends by age and gender, selection trends by situations depending on the presence or absence of passengers, selection trends including local specialties, timing and time This is a data conversion of the selection tendency. There are various selection trends such as a selection tendency of restaurants, a selection tendency of sightseeing spots, a selection tendency of rest places, and so on.
[0062]
Scenario data (10-2-3-4) is provided by the agent autonomously in any situation when the agent communicates with the driver. And a running execution condition that stipulates how to handle the execution of the scenario with respect to the running of the vehicle.
In the scenario data (10-2-3-4), image data to be displayed separately from the character (image data to be displayed on a scene display screen 54 (see FIG. 14) described later) is also stored.
[0063]
FIG. 10 shows the configuration of actual machine format scenario data.
The scenario data (10-2-3-4) is composed of a plurality of scenarios. The scenario data (10-2-3-4) is composed of data for managing them and data indicating the contents of each scenario.
The management data for the acquisition scenario includes information such as the expiry date of this scenario data, date of creation, creator, etc., and data for overall management of each scenario recorded in the scenario data (scenario number, Scenario name, priority (priority)), autonomous activation condition data of the scenario recorded in the scenario file, and the driver manually using the input device (5) in the scenario recorded in the scenario file Lists scenario list data that can be started.
[0064]
The data indicating the contents of each scenario includes management data for managing each scenario and scene data indicating the contents of individual scenes constituting the scenario.
The data for managing each scenario ("Management data for this scenario") includes information on the scenario, text information for creating a speech recognition dictionary used in this scenario, and the scene data that make up the scenario as a whole. The data for management is written.
[0065]
The scenario management data stores data for determining whether or not to perform standby processing. The standby process is a process for developing a standby scene that stands by in a standby state in the case of automatic activation.
This standby state is a state in which the execution of the scenario is waited until the user is ready to communicate with the agent by the agent performing the scenario execution notification and confirming whether or not to execute the scenario. For example, a small agent appears in the upper right corner of the map, speaks to the user, “If you touch me, I will introduce a recommended meal location,” and enters standby. When the user touches the agent, it enters the communication mode, and a scenario is developed that says, “There is a XX restaurant 2km ahead in the direction of travel. The
The user can select whether or not to shift to the standby state during scenario execution.
[0066]
The scene includes a normal scene in which a screen and sound are output, a branch scene in which no screen and sound are output, a clone scene in which a part of the normal scene is used, and a dummy scene.
A normal scene is composed of management data for managing a scene, screen configuration data, character motion data, various processing data, and development management data.
On the other hand, the branch scene is composed of management data for managing the scene and development management data.
[0067]
In the data for managing a scene, information on the scene and data for managing each data section belonging to the scene data are described.
In the screen configuration data, data (size, display position, etc.) of each part of the screen configuration displayed on the display device (2) in this scene is described.
The character action data includes instruction data for actions performed by the character in this scene and instruction data related to the content to be spoken. In the instruction data of the action, the instruction data is written in one of two types, that is, the scenario data that is directly instructed by the expression means of each character and the instruction that is instructed by the character to be expressed. A moving image corresponding to the action instructed by the instruction data is reproduced as the action of the character.
Various processing data includes information for controlling (processing) external devices in this scene, information for controlling navigation, instructions for executing other scenarios, timer setting information, and mental models that indicate character psychology. Information that changes the emotion element value is described.
The external device includes each device connected to the communication I / F unit (1-11), for example, a communication control device. The contents to be controlled include a process for making a call to a specific telephone number and a process for disconnecting a call.
The navigation control content includes, for example, setting this point as a destination.
As an instruction to change the emotional element value of the mental model, there are 1 to decrease the “friendship degree” of the long-term emotional element by 1 and “joy” of the short-term emotional element to 80.
[0068]
The development management data describes information such as whether to end the scenario, what is the next scene to proceed, or nothing to be developed when an event occurs in this scene. ing.
The event here refers to some action defined for proceeding to the next development of the scene. For example, when the character's dialogue has ended, the set time has elapsed, the driver has selected some answer to the question asked in this scene (for example, “Yes” to the question “Yes” or “No”) ), Etc. exist.
In addition to this event, the development can be changed according to the learning results.
For example, it can be used when the driver selects “Yes” for the question and the total number of uses is less than 10.
In addition to the learning results, the development can also be changed using the date and time, the mental state of the character using the mental model, driver information, and the like.
[0069]
FIG. 11 conceptually shows the contents of character data.
Character data (10-2-3-5) stores data of a plurality of characters, and can be selected from the input device (5) or the like according to the driver's preference.
The character data (10-2-3-5) includes character image data 102351, character voice data 102352, and character image selection data 102353 for each character A, B,.
[0070]
The character image data 102351 stores a still image representing the state of the character displayed in each scene specified by the scenario, an individual motion image (animation) representing the motion, and the like. For example, individual motion images are stored such as a movie that the character bows, a nod movie, and a right hand.
An individual motion image is a unit of a moving image that can express at least one complete meaning, such as nodding (affirmation) and shaking a head (no intention).
Each of these still images and individual motion images is provided with an image code.
The character image data 102351 functions as action expression storage means and image storage means.
The character (agent appearance) used as the character image data 102351 does not have to be a human (male, female) appearance. For example, the non-human agent may be the appearance of an animal itself, a robot-like appearance, a specific character appearance, or the like.
Also, the age of the agent does not need to be constant, and the learning function of the agent is to start with the appearance of a child, and then grow with the passage of time. It may change to the appearance of
[0071]
A scene (screen element) has one or a combination of two or more individual motion images, and is composed of individual motion images, sound, and other data.
FIG. 12 shows the contents of an individual action image representing the action in which the character raises the right hand.
As shown in FIG. 12, the individual operation image is composed of at least three types of moving images, that is, a starting moving image (a), a retained moving image (b), and an ending moving image (c). (A) to (c) represent representative images of the first, last, and halfway among the actually displayed moving images.
Individual actions of the character such as bowing are expressed by sequentially reproducing the three moving images (start moving image, retained moving image, and end moving image).
[0072]
As shown in FIG. 12A, the start moving image starts with an image (basic posture state image 12a1) representing the basic posture of the character and is an image (holding state image 12a4) in a predetermined posture holding state (predetermined state). finish.
In the starting moving image illustrated in FIG. 12A, an operation from the basic posture state to a smile with the right hand raised with the elbow bent is displayed.
[0073]
As shown in FIG. 12B, the retained moving image starts from the retained state image 12b1 and ends with the retained state image 12b4.
The retained moving image moves while retaining the content expressed by the character in the intermediate images 12b2 and 12b3 with the retained state image 12b1 as a reference, and then returns to the retained state image 12b4 as a reference.
In the retained moving image illustrated in FIG. 12B, the head tilts in the midway image 12b2 with respect to the retained state image 12b1, and the tilt of the head returns to the original in the midway image 12b3 so that the angle between the mouth and the wrist is It has changed.
The playback time of the individual motion image is adjusted according to the number of repetitions of the retained moving image. Further, when the shortest reproduction is performed on the individual operation image, the reproduction of the retained moving image is omitted. That is, the shortest moving image that passes through the holding state images 12a4 and 12c1 is executed by reproducing the ending moving image after the starting moving image.
[0074]
As shown in FIG. 2C, the end moving image starts with the holding state image 12c1 and ends with the basic posture state image 12c4. The end moving image is a moving image for returning to the basic posture state.
Note that the hold state image 12a4 at the end of the start moving image, the hold state images 12b1 and 12b4 at the start and end of the hold moving image, and the hold state image 12c1 at the start of the end moving image are the same image.
Further, the end moving image is the same as that reproduced from the starting moving image holding state image to the basic posture state.
[0075]
In this way, each individual motion image of the character starts and ends in the basic posture state, so that even if the individual motion image is composed of a plurality of individual motion images in each scene (screen element), the character motion can be continued.
[0076]
FIG. 13 conceptually displays the contents of the operation switching determination data (10-2-3-8) in the agent data (10-2-3).
Whether the operation switching judgment data (10-2-3-8) is transferred to the subsequent individual operation image in the immediate response mode or the quality mode corresponding to the interruption condition during the reproduction of the individual operation image and the interruption condition. It is a table that prescribes.
Each event (character movement end, etc.) defined in the “item” column shown in FIG. 13 corresponds to the interruption condition, the quality emphasis defined in the “motion switching determination” column is the quality mode, and the reaction emphasis is It corresponds to the quick response mode.
The agent device uses the operation switching judgment data (10-2-3-8) when the mode corresponding to the interruption condition is not specified in the scenario to be executed and when automatic selection is set in the scenario. Then the mode is decided.
[0077]
FIG. 11B conceptually shows a reproduction time table stored in the character image data 102351.
This playback time table stores a playback time when a start moving picture and an end moving picture in each individual motion image are played, and a playback time when a retained moving picture is played once.
The agent device reproduces one or a plurality of individual motion images in accordance with the definition of each scene data of the scenario, for example, the audio output time, etc., but the individual moving image is reproduced in accordance with the prescribed time. The playback time table is referred to in order to be displayed. For example, when the character bows while saying “Good morning, thank you again today”, it is assumed that the voice output time is calculated, for example, 6 seconds. In this case, from the playback time table, a playback time of 1.5 seconds is preferentially assigned to each of the start video and the end video, and a retained video is assigned to the remaining 3 seconds.
Since the time allocated to the retained moving image is 3 seconds and the reproduction time of the retained moving image is 1.5 seconds, N = 2 times, and the retained moving image is reproduced twice.
[0078]
Character voice data 102352 (FIG. 11 (a)) stores voice data for the agent to have a conversation with the driver according to the scene of the selected scenario.
The voice data of the conversation by the agent also stores voice data for the agent to ask questions for collecting driver information. As an example, "Hello", "I Regards", "And I", and the like are stored.
Each of these voices is accompanied by a voice code.
[0079]
The character image selection data 102353 is a conversion table in which image data representing the expression method (action) of each character is assigned to each display state.
Scenario data (10-2-3-4) defines the contents of each scene with a common display state that does not depend on the character type.
For this reason, the character image selection data 102353 is a conversion table for converting the display state of the scene expressed in common into image data that displays individual action contents for the character selected by the user. Function as.
[0080]
The driver information data (10-2-3-6) in FIG. 6 is not particularly shown, but is information related to the driver in order to make the agent's communication more suitable for the driver's wishes, hobbies and preferences. Used. This driver information data is also used as a scenario start condition and a scene transition condition.
Driver information data (10-2-3-6) contains the driver ID (identification information), name, age, gender, marriage (married or unmarried), child, for storing information for each driver. The driver basic data including the presence / absence, the number of persons, and the age, and hobby preference data are stored.
The hobby preference data is composed of major items such as sports, eating and drinking, and travel, and detailed items included in the concepts of these major items. For example, data such as whether or not to like baseball, whether or not to like soccer, whether or not to like golf is stored in the large-scale sports.
Driver information data (10-2-3-6) is created for each driver when there are a plurality of drivers who drive the vehicle. Then, the corresponding driver information is used by specifying the driver.
[0081]
In FIG. 6, learning item data and response data (10-2-3-7) are data for storing results learned by the agent by the driver's selection and response in communication with the agent.
Therefore, the learning item data and the response data (10-2-3-7) are stored / updated (learned) for each driver.
For example, as the usage status of the scenario, the previous selection result, the date and time of last use, the total number of times of use, and the like are stored.
In accordance with this learning content, for example, in a scenario in which greetings are made every time the navigation power is turned on, if it is within 5 minutes from the previous use, “I met you just before”, or conversely, for more than a month When it is open, it responds to “Long time no see”.
[0082]
FIG. 14 shows an example of a scene screen displayed on the display device (2) based on the scenario scene data.
The scene screen shown in FIG. 14 is a scene screen (scene number 0x0001) of a question scenario for asking a question from a driver in order to acquire a hobby preference (meal) that is uninputted driver information.
As shown in FIG. 14, the scene screen includes an agent display screen 51 on which an individual action image of the agent is displayed, a balloon screen 52 on which characters corresponding to the voice of the agent are displayed, a title screen 53, and a scene-specific screen. Are displayed on a scene display screen 54 on which image data (images of actual image data, answer selection buttons, etc.) are displayed.
The agent displayed on the agent display screen 51 is a character selected by the user or a default character.
[0083]
When the scenario driving unit (101-1) of the agent processing unit (101) starts a hobby preference (meal) question scenario, the scenario screen + image (10- 2-3-4), the scene screen is displayed on the display device (2), and the question voice corresponding to the question sentence is output from the voice output device (3).
On the scene screen of the question scenario in FIG. 14A, “Which genre do you like?” Is displayed on the balloon screen 52. And the audio | voice corresponding to the display of the balloon screen 52 is output from an audio | voice output apparatus (3).
Further, on the scene display screen 54 in the scene screen of FIG. 14A, “Japanese food”, “Western food”, “Chinese food”, and “None in particular” of the four answer selection buttons 54a are displayed.
Then, in accordance with the voice output of the character, an individual action image in which the character raises his right hand and points to the answer selection button 54a is reproduced and displayed.
[0084]
A plurality of scenes corresponding to the driver's answer are branched in the question scene for the driver. The branching of each scene and the identification of the subsequent scene are determined according to the driver's answer according to the development management data of each scene.
That is, when the driver selects the answer selection button “Japanese food” on the scene screen (scene number 0x0001) in FIG. 14A, the scenario driving unit (101-1) branches to the scene screen (b) corresponding to the answer. Is displayed. In this scene screen (b), the selected “Japanese food” is displayed on the title screen 53 and the balloon screen displays “I like Japanese food.” On the Japanese food scene screen after branching, The real Japanese food image 54 b read from the scenario data is displayed on the scene display screen 54. Then, the scenario driver (101-1) stores the driver's answer, for example, “Japanese food” as the driver information in the hobby preference data of the driver information data (10-2-3-6). It has become.
In this way, each scene image and sound specified in the scenario are sequentially displayed and output until the last scene, thereby completing the action of the agent in one scenario.
[0085]
FIG. 15 shows the transition of the scene screen according to the guidance scenario transmitted from the inn to the prospective guest for each scene.
This guidance scenario is composed of scene screens (a) to (f) among a plurality of scene screens.
The next scene screen is branched into 0x0004 and 0x0006 according to the selection result of the user with respect to the scene screen (c). Although not branched in the example of FIG. 15, the scene screen may be branched so that dishes corresponding to the selected dish type are displayed on the scene display screen 54 in the scene screen (d).
In this guidance scenario, a standby scene (s) that is displayed when standby processing is set by the user is set.
[0086]
Hereinafter, each action of the agent according to the reservation scenario will be described with reference to FIG.
The agent operations and screen displays described below corresponding to each scene screen are all displayed according to data, images, and instructions stored in the scenario data of the external scenario. Further, although described as an agent operation, the scenario driving unit (101-1) of the agent processing unit (101) actually performs the processing.
Moreover, although the image of the basic posture state is displayed on each scene screen, the individual operation image corresponding to each scene screen (a) to (f) is reproduced. For example, on the scene screen (a), an individual motion image that the agent bows is reproduced, and on the scene screen (d), an individual motion image (see FIG. 12) that raises the right hand while bending the elbow is reproduced.
[0087]
When the setting not to use the standby process is selected by the user, the reservation scenario is activated and developed from the scene having the number 0x0001.
On the other hand, when the setting to use standby processing is selected, the standby scene (s) in FIG. 15 is developed.
In the example displayed in FIG. 15, a screen immediately before the standby process is performed, that is, a map screen for route guidance by the navigation function is displayed. A small agent is displayed on the display screen (2) having a touch panel function in the upper right on the screen immediately before this. (Character display means)
Then, a notification of executing the scenario is made, such as “If you touch me,“ I will guide you to the reserved inn ”.” It also asks for permission to execute (confirmation of execution). If the voice data of the agent is defined in the scenario data, the voice is output.
Then, when an OK intention display by the driver is detected in accordance with the prompt for permission to execute, normal scenarios are sequentially developed from the first scene.
In the standby scene, when a device operation such as navigation is performed without a OK display and a certain time elapses, the execution of the scenario is terminated by a timer notification.
[0088]
When an OK indication is displayed in the standby scene, or when standby processing is not selected, the reservation scenario is activated.
That is, first, the scene screen with the number 0x0001 is displayed on the display device (2). In this scene, the agent of the character managed by the agent OS unit (101-8) appears on the agent display screen 51, bows and greets with voice. The content of the voice greeting is the same as the text displayed on the balloon screen 52.
The greeting by voice is performed on behalf of the inn by the agent, but it is expressed that the inn is a greeting from the inn by displaying a photo image of the general of the inn on the scene display screen 54. The image of the landlady is an image received and added as part of the external scenario, and is stored as actual image data of the scenario data (10-2-3-4).
The instruction for the action of the agent follows the instruction stored in the character action instruction data.
When the greeting by the agent ends, the scene transitions to the next scene 0x002.
[0089]
In the next scene 0x0002, an image of an open-air bath is displayed on the scene display screen 54. Then, the agent points to the picture of the outdoor bath, and the agent explains the specialty of the inn (that it is sold here) by voice and display of the balloon screen 52.
When the agent's story ends, the scene transitions to the next scene 0x0003, and an image of today's meal (an image of kaiseki cuisine) is displayed on the scene display screen 54. The agent explains the dish and whether or not this dish is acceptable. Question.
It is assumed that a timer setting time and a timer setting condition “set only during running” are defined in the scene data of scene 0x0003. In this case, counting by a timer is started at the start of the scene on condition that the vehicle is running. Whether or not the vehicle is traveling is determined to be stopped when the vehicle speed sensor (6-11) or the distance sensor (6-1-5) detects the vehicle speed v = 0, and the vehicle speed v ≠ 0 is detected. It is determined that the vehicle is traveling.
[0090]
If the user selects “Yes” as an answer to the question as to whether or not to change the displayed dish, the process branches to the scene 0x0004, and if “No” is selected, the process branches to the scene 0x0006.
On the other hand, if the user gives a timer notification (elapsed set time) without answering by voice or answering by selecting a selection button displayed on the screen within the timer set time, it is defined by the scene data of scene 0x0003. The scenario is terminated according to the transition condition at the time of timer notification.
In this way, even when there is no answer from the user, it is determined that no answer has been selected, and the process proceeds to the next scene (end in the example of FIG. 15) using no answer as a transition condition. It is possible to make the communication with the character converted closer to the communication between humans.
[0091]
In the scene 0x0004, a selectable list other than kaiseki cuisine is displayed on the scene display screen 54. The agent points to the list on the scene display screen 54 and asks which dishes are good.
When the user selects any one, the scene transitions to the scene 0x0005.
In the scene 0x0005, a list of the number of people to be changed from kaiseki cuisine is displayed on the scene display screen 54, and the agent points to this list and asks the number of people. When the user selects any one, the scene transitions to the scene 0x0006.
[0092]
In scene 0x0006, the inn appearance photograph image is displayed on the scene display screen 54, and the agent bows and greets.
Then, as a result of the selection made by the user, the agent transmits an answer result regarding meals to the third party (inn) that transmitted the external scenario being executed via the communication control unit in the case of the guidance scenario of FIG. .
In this way, when it is desired to acquire information about the user, the creator of the external scenario creates a scenario in which a scene of a question from which information to be acquired can be obtained is provided in the scenario and the response is transmitted by e-mail. To do. If it is necessary to send an answer, the e-mail address of the creator is included in the scenario data.
When the agent's story in the last scene (scene 0x0006 in FIG. 15) ends, the scenario ends.
[0093]
In this manner, the scenario driving unit (101-1) sequentially displays and outputs the individual operation images and sounds for each scene specified in the scenario up to the last scene.
When the activated scenario ends, the scenario driving unit (101-1) determines whether there is a request for activation of another scenario.
[0094]
Next, an operation for autonomously starting various scenarios executed by the scenario driving unit (101-1) will be described.
Autonomous activation determination unit (101-2) acquires status information such as the current position and time from agent OS unit (101-8) via agent I / F in order to acquire the current status information. Whether or not any of the autonomous activation conditions is satisfied is determined.
When the autonomous activation condition is satisfied, the autonomous activation determination unit (101-2) issues a scenario execution request message corresponding to the autonomous activation condition to the scenario driving unit (101-1). The scenario execution is started.
[0095]
FIG. 16 is a flowchart showing the flow of scenario execution processing.
Note that FIG. 16 shows a series of typical operations performed by each unit of the agent processing unit (101) and the overall processing unit (102) when a scenario is executed, and each unit performs independent processing. It is like that. That is, the independent process of each part is continued, and the typical flow shown in FIG. 16 is obtained.
Specifically, each part of the agent processing unit (101) and the entire processing unit (102) process the message when receiving the message, and wait for the next message when the processing is completed.
[0096]
Upon receiving the scenario execution request from the autonomous activation determination unit (101-2), the scenario driving unit (101-1) performs agent start preparation processing by securing and initializing work memory (step 505-1).
Then, the scenario driving unit (101-1) confirms whether the scenario execution request is a manual activation or an automatic activation (step 505-2). Manual activation is when the user selects scenario activation from the menu of the display device (2), and automatic activation is when the scenario autonomous activation condition is satisfied.
If the scenario execution request is manual activation, menu scenario request processing is performed (step 505-3). Thereafter, the process proceeds to the scenario data reading process (step 505-4).
[0097]
On the other hand, in the case of automatic activation, there is a scenario that is requested to be executed by satisfying the autonomous activation condition, so that the process directly proceeds to the scenario data reading process (step 505-4).
Next, the scenario driver (101-1) reads scenario data to be executed into the RAM (1-4) (step 505-4). When reading scenario data, if there are multiple scenarios to be executed (when multiple autonomous activation conditions are satisfied, manual activation request and automatic activation overlap, etc.), the scenario driver (101 -1) reads the scenario data with the highest priority by judging the priority specified in each scenario. If the priorities are the same, it is determined that the priorities are higher in the order in which the execution requests are received from the autonomous activation determination unit (101-2).
[0098]
When the reading of the scenario data is completed, the scenario driving unit (101-1) next performs a scenario start process (step 505-5).
In the scenario start process, the scenario driver (101-1) first performs an initialization process for starting the scenario. Furthermore, if the standby scene is selected, the scenario driving unit (1011) executes the standby scene described in FIG. 15 (s) and confirms the OK display, and the standby scene is selected. If not, the scenario start process is terminated.
[0099]
After the scenario start process (step 505-5), the scenario drive unit (101-1) executes a scene process for processing character drawing and sound according to the contents of the scenes constituting the scenario (step 505-6). ). Details of the scene processing will be described later with reference to FIG.
When the scene processing is completed, the scenario driving unit (101-1) checks whether the scenario is finished (step 505-7).
In the case of scenario termination, the scenario driver (101-1) performs scenario termination processing (505-8). In this scenario termination process, the learning unit (101-3) obtains an end ID indicating the termination method and stores it in the response data (10-2-3-7).
If the scenario is not finished (if the scenario still continues), the process returns to step 505-6 to repeat the scene process with the next scene, the next scene,.
[0100]
After scenario termination processing, the scenario driver (101-1) checks whether there is another scenario execution request (step 505-9). If there is another scenario execution request, the scenario drive unit (101-1) Returning to the data reading process (step 505-4), the same process is executed.
On the other hand, when there is no other scenario to be executed, the scenario driving unit (101-1) executes an agent termination process (step 505-10). That is, it notifies the agent OS unit (101-8) that the execution processing of all requested scenarios has been completed.
Thereafter, the screen displayed on the display device (2) returns to the normal navigation screen, but the subsequent processing is transferred to the overall processing unit (102) via the agent I / F.
[0101]
FIG. 17 is a flowchart showing the flow of scene processing (step 505-6).
The scenario driver (101-1) confirms the type of scene to be started in the scene processing (step 505-6-1), and in the case of a normal scene or a standby scene, the scene data analysis processing (step 505- In the case of a clone scene, the process proceeds to a process for requesting various processes (step 505-6-5).
Although not shown, when the scene is a dummy scene, the scenario driving unit (101-1) instructs the processing according to various processing data as background processing, and then determines the scene (step 505-13). Migrate to
On the other hand, in the case of a branch scene, since there are no various processes such as screen and audio output, the process proceeds to the scene determination process (step 505-6-13) in order to determine the next scene to be developed according to the additional conditions.
Here, the clone scene is a case where the same screen as the original scene (scene that ended immediately before) n is displayed depending on how a certain scene n ends, for example, when input is not performed within a set time. This is a scene in the case where a voice prompting input is output with the same screen.
In addition, a branch scene is a scene that is provided in front of the scene in order to make a transition to a specific scene and start the condition without making a screen display (branch). is there.
[0102]
When the scene to be started is a normal scene or a standby scene, the scenario driver (101-1) stores the data of the scene to be started in the RAM (in which the scenario data read in step 505-4 (FIG. 16) is placed ( 1-4), and analyze the screen configuration to be displayed, the action instruction of the character, etc. (step 505-6-2).
If the voice recognition dictionary exists as a result of the analysis, the scenario driving unit (101-1) sends a voice recognition dictionary setting (initialization) request specified in the scene data to the voice recognition unit (101-7). Notification is made (step 505-6-3).
In the case of a standby scene, an affirmative or negative speech recognition dictionary is set.
[0103]
Next, the scenario drive unit (101-1) is a screen having a screen configuration that determines each part of the screen to be drawn as a process for requesting the drawing / audio output unit (101-5) to draw the screen. Data creation processing is performed (step 505-6-4).
In the screen data creation processing of the screen configuration, for example, among the scene screens shown in FIG. 16, items related to the character such as the agent display screen 51 and the balloon screen 52 on which the character dialogue is displayed are determined.
[0104]
The scenario drive unit (101-1) instructs various processes (step 505-6-5).
The various processing instructions include processing to a navigation system, an externally connected device, or the like, or time measurement request processing when a timer is set.
In addition, the scenario driving unit (101-1) is set when various instructions are set for the scene as long-term emotion elements or short-term emotion elements. The mental psychology department (101) changes the long-term emotional element 10-2-3-1a and the short-term emotional element 10-2-3-1c of the mental model data (10-2-3-1) -4).
The character psychology unit (101-4) changes the value of each emotion element according to the change value specified in the scene according to the instruction. That is, if it is a long-term emotional element, the indicated change value is added or subtracted, and if it is a short-term emotional element, the value of the short-term emotional element is changed to the indicated change value and other short-term emotional elements Set the value to zero.
In this way, long-term emotional changes or short-term emotional changes in the agent device are pre-determined in the agent device (long-term emotional change condition 10-2-3-1b, short-term emotional change condition 10-2 -3-1d), it is possible to change according to the emotion element change instruction set in the scenario. For this reason, it becomes possible to realize the agent's emotional change according to the intention of the scenario creator.
[0105]
Next, the scenario driving unit (101-1) determines the character action of the scene and the speech such as dialogue, and makes a request for drawing / sound output of the determined character to the drawing / sound output unit. Accordingly, the drawing voice output unit performs character drawing / voice output processing (step 505-6-6).
The creation of the character drawing data by the scenario drive unit (101-1) is the same as the drawing data creation processing of the screen configuration shown in FIG. 16 except that the scenario part to be created is different from the part of the screen configuration or the part related to the character. To be done. In the case of creating the character drawing data, the voice of the character corresponding to the dialogue displayed on the balloon screen 52 and the voice data regarding the sound effect are also specified.
When creating the character drawing data (including voice data), the scenario driving unit (101-1) makes a character drawing request based on the created character drawing data to the drawing / voice output unit (101-5).
The drawing / sound output unit (101-5) performs character drawing / sound output processing in accordance with a drawing request from the scenario driving unit (101-1). By the character drawing / speech output process, a scene is developed in which the character bows, points in the right or left direction by hand, or talks to the user.
In the case of a standby scene, a character smaller than normal (FIG. 15A) as illustrated in FIG. 15S is displayed.
[0106]
FIG. 18 is a flowchart showing character drawing / sound output processing by the drawing / sound output unit (101-5).
Upon receiving a character motion instruction request or drawing data from the scenario driving unit (101-1), the drawing / sound output unit (101-5) receives a motion instruction content analysis process (steps 505-6-6-1 to 505). -6-6-8), operation reproduction processing (step 505-6-6-9), request end reply (step 505-6-6-10).
[0107]
In the action instruction content analysis process, the drawing / sound output unit (101-5) first determines whether or not it is a unified (common) action instruction that does not depend on the type of the received drawing instruction content character (step 505-6). -6-1) If the drawing instruction is instructed by the expression method specified for each character (direct instruction: see FIG. 37 (b)), the process proceeds to the action reproduction process (step 505-6-6-9). To do.
[0108]
If the drawing instruction is a unified display form that does not depend on characters, the drawing / audio output unit (101-5) converts the instruction content.
In the conversion, first, the character type set at that time is acquired from the agent OS section (101-8) (step 505-6-6-2).
Next, the drawing / sound output unit (101-5) displays a conversion table (character image selection data) 102353 in which a correspondence table of unified operation instructions (display state numbers) and operation instruction contents (image codes) for each character is written. Is obtained from the character data (10-2-3-5) of the external storage device (10).
Then, the drawing / audio output unit (101-5) acquires the action instruction content of the character to be operated based on the conversion table, that is, the image code of the character corresponding to the display state number of the scene data (step 505-6). -6-4).
[0109]
The drawing / sound output unit (101-5), when the acquired action instruction information is set to automatically select the action instruction of the character (step 505-6-6-5; Y), Further, the following processing is performed.
That is, the drawing / speech output unit (101-5) first sets the automatic operation condition information, such as time, place, driver information, agent mental model, etc., to each information included in the agent processing unit such as the agent OS unit. Is acquired from the processing unit (101-8) that manages (step 505-6-6-6).
Next, automatic selection from agent data (10-2-3) that has already been read into the RAM when the power is turned on, including selection condition information for uniform operation instructions such as time, and uniform operation instructions for characters A table is acquired (step 505-6-6-7).
Then, the drawing / sound output unit (101-5) acquires the unified operation instruction based on the selection condition information of the unified operation instruction such as time and the operation automatic selection table. Based on the unified operation instruction, the operation instruction is determined by referring to the conversion table 102353 acquired in step 505-6-6-3 to acquire the operation instruction content (image code) (step 505-6- 6-8).
[0110]
Next, the drawing / sound output unit (101-5) performs operation reproduction processing (step 505-6-6-9).
That is, the drawing / sound output unit (101-5) performs character image data 102351 of the selected character in the character data (10-2-3-5) based on the action instruction content (image code) of the character. The individual motion image to be reproduced is acquired from (see FIG. 11).
If the action instruction request for the character is to synchronize and output the character speech to the character action, the voice data is acquired from the character voice data 102352 of the character data (10-2-3-5).
[0111]
The drawing / audio output unit (101-5) obtains the output time t1 of the audio data when outputting the audio data, and is specified in the scene data when there is no output of the audio data. Time t1 is acquired. Hereinafter, since the time t1 is the total time allocated for playback of individual moving images, it is referred to as a total allocated time.
The drawing / audio output unit (101-5) is assigned to the line or time specified in the scene data (it is specified to be played during the output of the audio of the line or within the specified time). The playback time of the individual motion image is acquired from the playback time table (see FIG. 11B).
[0112]
The drawing / sound output unit (101-5) determines the number of times N of retained moving images to be played in each individual motion image from Equation (1) from the obtained playback time of each individual motion image and the acquired allotted time t1. To do.
In this equation, the total allocated time is t1, the playback time of the start video is t2, the playback time of the retained video is t3, and the playback time of the end video is t4. Further, the total values of the reproduction times of all start videos, all retained videos, and all end videos are Σt2, Σt3, and Σt4, respectively.
In this case, time Σt2 is assigned to all start videos, time Σt4 is assigned to all end videos, and time (t1- (Σt2 + Σt4)) is assigned to all retained videos.
Thus, time is preferentially assigned to the start video and end video of each individual motion image, and the remaining time is assigned to each retained video.
[0113]
[Formula 1]
n = (t1− (Σt2 + Σt4)) / Σt3 (remainder time t5) (1)
[0114]
Then, the drawing / sound output unit (101-5) calculates n and the remainder time t5 from the formula (1), and determines the number of times to play each retained moving image as follows.
(A) When t5 = 0
Number of playbacks of all retained videos N = n
(B) When t5 ≠ 0
Number of playback times from the first retained video to the m−1th retained video N = n
Number of plays from the m-th retained video to the last retained video N = n + 1
However, m is the maximum value of p satisfying {Σt3} ≧ t5, where {Σt3} is the total reproduction time from the p-th retained moving image to the last retained moving image.
For example, when 7 seconds are allocated to the retained moving images from a to d (each reproduction time of 1 second), n = 6/4 = 1 and remainder 2, so that the number of reproductions of a and b is 1 time. , C and d are reproduced twice.
[0115]
In the present embodiment, the playback of the individual motion image ends at the same time or after the all allocated time t1, but the playback of the individual motion can be reproduced by setting the number of times N of the retained moving image to be played back uniformly. You may make it complete | finish before a regulation time. In this case, the basic (standing) retained moving image shown in FIG. 11B is repeatedly reproduced until the audio output is completed or until the specified time elapses.
Further, when the total allocation time t1 is shorter than the sum of both playback times of the start video and the end video, only the start video and the end video are played.
[0116]
FIG. 19 shows a flow that can be transferred between the start moving image, the retained moving image, and the end moving image when the individual operation image is reproduced.
As shown in FIG. 19, playback of each moving image is continued through the holding state screen.
(1) Video following the start video
The start moving image can be transferred to the following two moving images.
{Circle around (1)} When the number of times the retained moving image is played back N ≠ 0, the process proceeds to the retained moving image.
{Circle over (2)} When the number of times the retained moving image is played back N = 0, the process proceeds to the end moving image.
(2) Video that follows the retained video
The retained moving image can be transferred to the following two moving images.
{Circle around (3)} When the number of playback times of the retained video N> 2, the process shifts to the retained video.
{Circle around (4)} When the number of times the retained moving image is played back N = 1, the process proceeds to the end moving image.
(3) Video following the end video
The end moving image does not move to the moving image in the same individual operation image, but moves to the starting moving image of the subsequent individual moving image.
[0117]
Next, a case where the interruption condition is satisfied during the reproduction of the individual operation image will be described.
The scenario driving unit (101-1) monitors whether or not the interruption condition (condition described in the item column) defined in FIG. 13 is satisfied during reproduction of the individual operation image.
When any of the interruption conditions is satisfied, the scenario driver (101-1) draws the transition mode (quality mode, quick response mode) for transitioning to the next individual motion image, and the interruption event / audio. Supply to the output unit (101-5).
When the transition mode is defined for the individual motion image being reproduced, the transition mode is determined as the defined transition mode. When the automatic mode is specified for the individual operation image and when the transition mode is not specified, the operation switching judgment data (10-2-3-8) (see FIGS. 6 and 13) is used. Thus, the transition mode corresponding to the interruption condition is determined.
As described above, the mode determination means is configured by determining the transition mode by the scenario driving unit (101-1).
[0118]
When an interruption event is supplied from the scenario drive unit (101-1), the drawing / audio output unit (101-5), depending on the transition mode (quality mode, quick response mode), , Holding movie, end movie) are interrupted, and the subsequent individual motion image is transferred (executed).
[0119]
FIG. 20 shows a reproduction state after interruption in the quality mode.
As shown in FIG. 20, in the quality mode, the occurrence of the interruption event differs depending on which moving image is being reproduced.
(1) When starting video playback
As shown in FIG. 20A, when an interruption event occurs during the playback of the start video, after the start video is finished playing, the retained video is played once, and after the end video is played, the subsequent individual operations Move to image.
If an interruption event occurs during playback of the start video, after the playback of the start video is finished, the end video is played without playing the retained video, and the subsequent individual motion image is transferred. Good. In this case, since the hold state image is displayed at the end of the start video and the start of the end video, there is no gap when moving to both videos, and the subsequent individual motion image can be played back as soon as the playback of the hold video is omitted. it can.
[0120]
(2) During playback of retained video
As shown in FIG. 20B, when an interruption event occurs during the playback of the retained video, the end video is played after the retained video being played is played to the end, and then the subsequent individual motion image Migrate to
When the retained moving image is being reproduced, if the p-th reproduction is being performed (p <N) with respect to the N times of reproduction determined from the above equation (1), the retained moving image is omitted Np times. . Therefore, it is possible to ensure a certain degree of responsiveness without gaps between moving images while maintaining quality.
(3) End video playback
As shown in FIG. 20C, when an interruption event occurs during the reproduction of the end moving image, the end moving image is reproduced as it is, and then the subsequent individual operation image is transferred.
[0121]
In this way, in the quality mode, there is no gap between the moving images being played back, the interrupted individual motion image ends in the basic posture state, and the subsequent individual motion images start from the basic posture state. And the transition to the subsequent individual operation image is possible.
[0122]
FIG. 21 shows a reproduction state after interruption in the quick response mode.
As shown in FIG. 21, in the case of the quick response mode, the moving image being reproduced is interrupted and the subsequent individual motion image is transferred regardless of the occurrence time of the interruption event.
In the case of the quick response mode, there is a slight gap when moving to the subsequent individual video because it is interrupted when the interrupt event occurs, but for example, the individual operation of the subsequent with good response without causing unnecessary waiting for the agent device user. Images can be played back.
[0123]
As described above, when there is an interruption event, the reproduction data of the individual motion image determined by the drawing / audio output unit (101-5) is transmitted to the agent OS unit (101-8) when there is no interruption event, and the external I / F unit (101 -9) is transmitted to the overall processing unit (102), and is transmitted to the drawing processor (1-6) through the processing unit for instructing the drawing processor (1-6) in the overall processing unit (102). It is sequentially reproduced and displayed on the display device (2).
The acquired voice data is transmitted to the agent OS unit (101-8) and from the external I / F unit (101-9) to the overall processing unit (102), and the voice processor (1- The signal is transmitted to the audio processor (1-8) through the processing unit for instructing 8), and the audio output control signal is converted into an analog signal and output to the audio output device (3).
[0124]
Next, the drawing / sound output unit (101-5) completes the character drawing / sound output processing for the requested scene after performing all the motion processing of the character requested by the scenario driving unit (101-1). The scenario driver (101-1) is notified (FIG. 18, Step 505-6-6-10), and the process is terminated.
If the content of the request is to synchronize and output the character speech, the scenario driver (101-1) is notified of the completion of the character speech output processing when all the speech output is completed.
[0125]
After the drawing / sound output unit (101-5) notifies the end of the character drawing / sound output process of the scene, the scenario drive unit (101-1) is instructed about the speech recognition in the scene data being processed. (FIG. 17, step 505-6-7). If there is no instruction, the process proceeds to step 505-6-9. If there is an instruction, the voice recognition specified in the scene data is detected. Voice recognition processing is performed using the dictionary for the above (step 505-6-8).
[0126]
The scenario driving unit (101-1) notifies the user input from the agent OS unit (101-8) when there is no voice recognition instruction and after the voice recognition process (step 505-6-8) is completed. When it is received, it is confirmed what the input is (step 505-6-9), and processing corresponding to the input is performed.
As described above, since each process is executed independently, when an input is notified from the agent OS unit (101-8) even during the voice recognition process (step 505-6-8). The processing corresponding to the input is executed in parallel. Therefore, when the user selects the voice recognition selection button and is notified from the agent OS unit (101-8) during the voice recognition process, the next process (step 505-6) is performed regardless of the processing stage of the voice recognition process. -9) is executed.
For example, in the case of a standby scene, it is possible to determine whether or not the driver has indicated OK (affirmation) based on whether or not the screen of the display device has been touched. The next process (step 505-6-9) is executed regardless of the processing stage of the voice recognition process.
[0127]
When the scenario driving unit (101-1) receives, for example, an input related to the movement of the cursor from the agent OS unit (101-8), the scenario driving unit (101-1) moves the cursor to process a screen drawing request (screen scroll request). (Step 505-6-10).
When the user selects any answer selection button 54a displayed on the scene display screen 54 (see FIG. 14), it is determined which item has been selected (step 505-6-11).
Note that, as described above, the processes in FIGS. 16 and 17 show an example of the scenario process. In practice, each unit independently performs an individual process. For this reason, although not shown in FIG. 17, when voice recognition start or stop is input, confirmation of user input such as requesting the voice recognition unit to start or stop voice recognition processing is performed (step 505-6). There are other processes after -9). In addition, even before the end of the character drawing / audio output process of the scene is notified from the drawing / audio output unit (101-5), that is, before the operation of the instructed character is completed, It is also possible to perform steps 505-6-9).
[0128]
Next, the scenario driving unit (101-1) refers to the scenario data development management data (see FIG. 10) in the development judgment process (step 505-6-12) from the judgment result of the selected item. Determine next deployment.
When there is no next development, the scenario driving unit (101-1) returns to the user input determination without performing any processing.
[0129]
On the other hand, if there is a next development, and if the scene determined in step 505-6-1 is a branch scene, the scenario driving unit (101-1) determines the scene to be developed next. (Step 505-6-1).
In the scene determination process, the scenario drive unit (101-1) acquires the transition condition set for the currently processed scene from the scenario data development management data (see FIG. 10).
Then, the scenario drive unit (101-1) displays the target state change (defined in FIG. 25) or the target state corresponding to the acquired transition condition (development condition or additional condition) as the character psychology unit (101-4). And from the agent OS section (101-8).
Then, a normal scene, a clone scene, or a branch scene to be developed next, which is defined in the development management data with respect to the transition condition that the received state change or target state satisfies, is determined.
[0130]
The scenario drive unit (101-1) determines the next scene to be developed by the scene determination process, and then proceeds to the scene end process (step 505-6-14) to proceed to the next development.
In the scene end process (step 505-6-14), if there is a process requested by the scenario drive unit (101-1) to another processing unit, a cancellation request is made (for example, a request for a speech recognition process is made). If so, request to stop the recognition process) and return. Upon return, the process proceeds to scenario end determination (step 505-7) in FIG.
[0131]
Next, the configuration and operation of a scenario creation device in which a user or a third party creates a unique scenario will be described.
FIG. 22 shows the configuration of the scenario creation device.
The scenario creation device includes a control unit (200), an input device (210), an output device (220), a communication control device (230), a storage device (240), a storage medium drive device (250), Input / output I / F (260). These devices are connected by a bus line such as a data bus or a control bus.
[0132]
The control unit (200) controls the entire scenario creation device.
The scenario creation device can execute not only the scenario editing program but also other programs (for example, a word processor and a spreadsheet). The control unit (200) includes a CPU (200-1) and a memory (200-2).
The CPU (200-1) is a processor that executes various arithmetic processes.
The memory (200-2) is used as a working memory when the CPU (200-1) executes various arithmetic processes.
The CPU (200-1) can write and erase programs and data in the memory (200-2).
In the memory (200-2) in the present embodiment, an area for the CPU (200-1) to create, edit, and store scenario data according to a scenario editor (scenario editing program) can be secured.
[0133]
The input device (210) is a device for inputting characters, numbers, and other information to the scenario creation device, and is composed of, for example, a keyboard and a mouse.
The keyboard is an input device for inputting mainly kana and English characters.
The keyboard is used, for example, when a user inputs a login ID or password for logging in to the scenario creation device, or when a sentence to be subjected to speech synthesis or speech recognition is input at the time of scenario creation.
A mouse is a pointing device.
When operating a scenario creation device using a GUI (Graphical User Interface) or the like, an input device used for inputting predetermined information by clicking a button or icon displayed on the display device It is.
[0134]
The output device (220) is, for example, a display device or a printing device.
As the display device, for example, a CRT display, a liquid crystal display, a plasma display or the like is used.
Various screens such as a main screen for creating a scenario and a screen for selecting a screen configuration in each scene are displayed on the display device. In addition, selected information and input information are displayed on each screen.
As the printing apparatus, for example, various printer apparatuses such as an ink jet printer, a laser printer, a thermal transfer printer, and a dot printer are used.
[0135]
The communication control device (230) is a device for transmitting and receiving various data and programs to and from the outside, and a modem, a terminal adapter, and other devices are used.
The communication control device (230) is configured to be connectable to, for example, the Internet or a LAN (Local Area Network). The communication control device (230) exchanges signals and data by communication with other terminal devices or server devices connected to these networks, thereby transmitting scenario data created by the device, or by a third party. The created scenario data can be received (downloaded), and further, data necessary for creating the scenario data can be acquired.
The communication control device (230) is controlled by the CPU (200-1), and performs transmission and reception of signals and data with these terminal devices and server devices according to a predetermined protocol such as TCP / IP.
[0136]
The storage device (240) includes a readable / writable storage medium and a drive device for reading / writing programs and data from / to the storage medium.
A hard disk is mainly used as the storage medium, but it can also be configured by other readable / writable storage media such as a magneto-optical disk, a magnetic disk, and a semiconductor memory.
The storage device (240) stores a scenario editing program (240-1), scenario editing data (240-2), and other program data (240-3). Other programs include, for example, a communication program that controls the communication control device (230) and maintains communication with the terminal device and server device connected to the scenario creation device and the network, and scenario creation such as memory management and input / output management An OS (Operating System), which is basic software for operating the apparatus, is also stored in the storage device (240).
[0137]
The storage medium driving device (250) is a driving device for driving a removable storage medium to read / write data. Examples of the removable storage medium include a magneto-optical disk, a magnetic disk, a magnetic tape, IC cards, a paper tape punched with data, and a CD-ROM.
In the present embodiment, scenario data created and edited by the scenario creation device (a form used by the agent device) is mainly written in IC cards.
The scenario creation device acquires a scenario from the storage medium storing the scenario data by driving the storage medium by the storage medium drive device (250) or stores the created scenario data from the storage medium drive device. Can be stored.
[0138]
The input / output I / F (260) is configured by, for example, a serial interface or another standard interface.
The function of the scenario creation device can be expanded by connecting an external device corresponding to the interface to the input / output I / F (260). Examples of such external devices include a storage device such as a hard disk, a communication control device, a speaker, and a microphone.
[0139]
Next, the configuration of the scenario editing program (240-1) and scenario editing data (240-2) will be described.
FIG. 23 conceptually shows the configuration of the scenario editing program and data.
The scenario editing program (240-1) includes a scenario editor (240-1-1), a scenario compiler (240-1-2), and a DB editing tool (240-1-3).
The scenario edit data (240-2) includes the common definition DB (240-2-1), local definition DB (240-2-2), and SCE scenario data (240-2-3) created with the scenario editor. And actual machine format (NAV format) scenario data (240-2-4) converted by the scenario compiler.
The scenario editor (240-1-1) is an application program that creates scenario data.
[0140]
The scenario compiler (240-1-2) converts the SCE scenario data (240-2-3) created by the scenario editor (240-1-1) into actual machine scenario data (240- It is an application program that converts to 2-4) and functions as a conversion means.
FIG. 24 conceptually shows data format conversion.
As shown in FIG. 24, the scenario compiler (240-1-2) converts one or more SCE format scenario data (240-2-3) into one actual machine format (NAV format) scenario data (240 Convert to -2-4).
[0141]
The DB editing tool (240-1-3) is an application program for editing / updating data stored in the common definition DB (240-2-1).
The common definition DB (240-2-1) stores definition data for creating scenario data. The common definition DB (240-2-1) stores automatic start condition items, development condition items that specify scene development conditions, additional condition items (transition conditions), a character display state instruction table, and the like. This common definition DB (240-2-1) may exist not on the storage device of the scenario creation device but on a server connected by a local area network (LAN). By doing so, it becomes possible for each scenario creating apparatus connected by a local area network (LAN) to create scenario data using a common common definition DB (240-2-1).
The local definition DB (240-2-2) stores the screen configuration defined by the scenario creator while creating scenario data.
[0142]
The SCE format scenario data (240-2-3) is data created by the scenario editor (240-1-1).
The actual machine format (NAV format) scenario data (240-2-4) is converted from the SCE format scenario data (240-2-3) to the data format for use on the agent device by the scenario compiler (240-1-3). Data.
[0143]
Examples of items that can be edited with the scenario editing program are shown below. These are stored in the common definition DB (240-2-1).
FIG. 25 shows a development condition item table in which transition conditions (development conditions) for development from the normal scene to the next scene (normal scene, branch scene) are stored. . In the development condition item table, as shown in FIG. 25, state changes of various objects such as character action end, scenario interruption, and “Yes” selected as a user response are defined.
A plurality of expansion condition items can be set for one normal scene, but one expansion condition item can be set between a normal scene and another normal scene and between a normal scene and a branch scene. .
The expansion condition item table is stored in the common definition DB (240-2-1).
Each item of the development condition item is read out when creating the development configuration of each scene, and is displayed in a list in the branch condition designation window. Select the expansion condition item from the list displayed, and if it is not stored in the table, define the expansion condition separately using the DB editing tool (240-1-3) and add it to the scene. A deployment configuration is created.
It is possible to create a scene having a plurality of transition destinations (branched into a plurality of branches) by repeatedly selecting the expansion condition item for one normal scene and then setting a plurality of normal scenes or branch scenes to be expanded thereafter. it can.
[0144]
FIG. 26 and FIG. 27 conceptually show part of the contents of a unified action instruction table that does not depend on characters. Hereinafter, character setting means is formed from the description of FIGS. 26 and 27.
This table defines a common action display state regardless of the type of character, and is categorized for each content to be expressed by the character.
There are a plurality of unified motion instruction tables that do not depend on characters. In this embodiment, the display state instruction tables for work state (FIG. 26), mental state (FIG. 27), TPO state, growth state, and scale state are provided. Exists.
As shown in FIGS. 26 and 27, each display state instruction table has a plurality of tree structures, and the shape and classification name of each tree structure are displayed in the character action state instruction edit window described later. It has become.
[0145]
As shown in FIGS. 26 and 27, a state instruction number is assigned to each item at the end of the tree of the display state instruction table. This state instruction number corresponds to the state instruction number (see FIG. 9) of the character image selection data (conversion table) 102353 of the agent device 1.
[0146]
In these tables, as shown in FIG. 26 and FIG. 27, the system (agent device) indicates the level defined in the lower layer for the display state (carefully, normally, strong, medium, weak, etc.). The item “automatic” is automatically selected.
For the scene selected as automatic, the agent device determines which level of the display state of the character to use based on the character psychological state, date, time, etc., and selects and executes one of the display states. It will be.
[0147]
In the common definition DB (240-2-1), in addition, data for speech recognition used for speech recognition, data used for character motion instructions (separate speech instruction data also exists), and set for each scene Character image data and character dialogue data for previewing and confirming the designated instruction, a conversion table to each character representation method for a unified instruction independent of the character, each part data displayed on the display device (2), and Screen configuration data describing how to arrange the items and items that can be selected as processing contents in the scene, for example, actions that can be processed by the agent, on / off of the audio device, channel selection, air conditioner device Common definition of various process content item data such as on / off, temperature setting, destination setting to be supplied to the overall processing unit (102) D Stored in (240-2-1).
The common definition DB (240-2-1) also contains items related to startup conditions for editing automatic startup conditions and additional condition items (items set in branch scenes) for diversifying scenario development. Stored.
All of these definition data can be changed and added using the DB editing tool (240-1-3) in the same manner as each definition data.
[0148]
Character image data for previewing and confirming instructions set in each scene stores image data of various characters stored in the agent device.
Note that the user can also store character image data and conversion tables for other characters from the agent device in the common definition DB of the scenario creation device via the IC card 7 or the server 3.
[0149]
Next, each operation of scenario creation by the scenario creation device configured as described above will be described according to the transition of the screen.
FIG. 28 shows the configuration of the main window displayed on the display device when the scenario editor (240-1-1) is activated.
As shown in FIG. 28, the main window includes a scene screen 301 on which a scene screen being created (a scene screen (see FIG. 14) displayed on the display device (2) of the agent device 1) is displayed. A setting screen 303 on which setting items to be set are displayed, a scene development screen 305 in which a scene development configuration (branch state) is displayed by a tree structure of scene icons 307 representing each scene, and a processing icon display unit Has been.
[0150]
When the scenario editor (240-1-1) is activated, a start point 308 is displayed on the scene development screen 305 of the main window. When this start point 308 is selected, scenario properties can be edited. The selection is made by, for example, matching the point position by the mouse cursor to the corresponding location and double-clicking the mouse.
A screen configuration change button 309 is a button for selecting a screen configuration to be displayed, and a sound effect setting button 110 is a button for displaying a screen for setting a sound effect for each scene of the scenario.
When the agent display screen 311 is selected, an agent (character) action editing screen is displayed.
The dialogue editing button 313 is a button for editing a dialogue instruction of the character. When the button part and the background speech recognition dictionary setting 315 are selected, the speech recognition dictionary to be used can be edited. The name of the word to be recognized is displayed on the scene screen when the one displayed with the mark of the answer selection button 315a (54a) on the scene screen 301 is selected, and when the one 315b to be recognized in the background is selected, the target of speech recognition is selected. However, the name of the recognized word is not displayed.
[0151]
The timer setting button 317 is a button for setting and changing timer setting information.
In the control instruction edit 319 for an external device or the like, a control instruction for an external device or the like (including a navigation) is set.
The voice recognition start control instruction 320a sets a voice recognition instruction that defines how to start voice recognition when voice recognition is performed in the scene being created. As the voice recognition instruction, one of “automatically start”, “not automatically start”, and “judgment by agent device (on-vehicle device)” can be selected. Yes.
In the callback control instruction 320b, an instruction as to whether or not to perform a callback for confirming the result of voice recognition is set. As a callback instruction, one of “call back”, “do not call back”, or “agent device determines (automatic)” that determines whether the agent device determines the situation and determines whether to call back or not. Can be selected.
The AMM change setting change button 321 is a button for changing each long-term emotion element of the agent in the scene being created, as will be described later.
[0152]
The scene development display unit 322 displays a development condition item and an additional condition item defined for a scene designated (active) on the scene development screen, and a scene to be continued (connected) when the condition item is satisfied. Is displayed.
On the left side of the scene development display unit 322, as shown in FIG. 28, the development conditions (see FIG. 24) and the classification of additional conditions (see FIG. 25) set in the active scene are displayed in a tree structure. . In addition, on the right side of the scene development display section, the contents of the transition condition and the connection destination (the scene number of the development destination) are displayed.
[0153]
The processing icon display section includes a scene creation button 323, a branch scene creation button 324, a dummy scene creation section 325, an alias scene designation button 326, an end ID button 327, a link ID button 328, an insert button 329, an up / down change button 330, A scene playback button 331, a build button 332, and other processing buttons are displayed.
[0154]
The scene creation button 323 is used when a new normal scene is created.
When the scene creation button 323 is clicked, the scenario flow can be edited (the next scene is created). When the scene creation button 323 is clicked, a normal scene to be developed next to the currently selected scene can be created.
By branching the flow of the scenario with the scene creation button 323, a development configuration for each normal scene is created. For example, in the scene development configuration displayed on the scene development screen 305 in FIG. 28, when the scene creation button 323 is clicked in a state where the normal scene icon 5 is selected (actively displayed), the scene 5 is displayed. The icon of the subsequent scene is displayed on the lower layer side, and the normal scenes 7, 8,... Developed after the scene 5 are branched and created by clicking a plurality of times (a plurality of creation operations are performed).
[0155]
That is, by clicking the scene creation button 323 with the scene m (normal scene or branch scene) selected, the next normal scene m1 following the scene m is created. When the scene creation button 323 is clicked with the scene m selected again, the next normal scene m2 following the scene m is created by branching in parallel with the normal scene m1. Similarly, when the scene creation button 323 is clicked with the scene m selected again, the normal scene m3 is created.
If the scene following the created normal scene m1 is to be further developed, the next normal scene m1-1 following the normal scene m1 is clicked by clicking the scene creation button 323 with the normal scene m1 selected. Created. When creating another scene that branches off from the normal scene m1, the scene m1 is selected again and the scene creation button 323 is clicked to create a scene m1-2 following the scene m1.
Furthermore, if the scene creation button 323 is clicked while the created normal scene m1-1 is activated, a normal scene m1-1-1 following the normal scene m1-1 is created.
[0156]
A branch scene creation button 324 is used when a new branch scene is created.
When the branch scene creation button 324 is clicked, the scenario flow can be edited (create the next branch scene). When the branch scene creation button 324 is clicked, it is possible to create a branch scene to be developed next to the currently selected scene.
By using the branch scene creation button 324 to branch under a plurality of additional conditions following the development condition, a development configuration for each normal scene by a logical operation (logical sum, logical product) of the transition conditions can be created.
For example, in the scene development configuration displayed on the scene development screen 305 in FIG. 28, when developing from the normal scene 1 to the normal scene 2, by inserting the branch scene 4 and the branch scene 6 between the two normal scenes, When the state change defined in the normal scene 1 is performed and both the branch condition 4 and the branch condition 6 are satisfied, the normal scene 2 is developed following the normal scene 1. In this way, by using a plurality of branch scenes, it is possible to create scenarios that perform various developments with various condition settings.
[0157]
In FIG. 28, the branch scene is denoted by reference numeral 333. The branch scene can be created following a normal scene or a branch scene. For example, a branch scene 4 is created by selecting the branch scene creation button 324 with the normal scene 1 on the scene development screen 305 in FIG. 28 activated, and a branch scene creation button with this branch scene activated. A branch scene 6 is created by selecting 324.
The method for creating the branch scene hierarchically is the same as the method for creating the normal scene described above.
[0158]
The dummy scene creation button 325 is a button for creating a dummy scene.
The dummy scene is composed of management data for managing the scene, various processing data, and development management data, and each week process such as destination setting processing in the navigation function and storage processing of the agent learning result is defined. The
[0159]
The alias scene designation button 326 is a button for shifting from an active scene (normal scene, branch scene) to another created scene (normal scene, branch scene).
For example, when the scene scene 5 on the scene development screen 305 has the branch scene 6 active and the alias scene designation button 326 is selected and then the scene screen 5 to be transferred is selected, the alias indicating that the scene screen 5 is transferred to. A screen 334 is displayed.
[0160]
The end ID button 327 is a button for creating an end ID for specifying the end position of the scenario.
When the end ID button 327 is clicked, a scenario end position 340 can be created. An end number is assigned as an end ID to the end position 340 of each created scenario.
When the end position 340 of the created scenario is selected, an end property editing screen is displayed, and each short-term emotion element of the agent can be set and changed on this screen.
[0161]
The link ID button 328 is a button for creating a link ID for linking another scenario to the scene.
The insert button 329 is a button for inserting a normal scene, a branch scene, and a dummy scene before each active scene (normal scene, branch scene) or end mark. Select a button for each scene type.
The up / down replacement button 330 is a button for changing the vertical position of the scene, and includes a button for moving the active scene up and a button for moving down the active scene.
The scene playback button 331 is a button for playing back an active normal scene.
The build button 332 is a button for compiling the created scenario into an actual machine format (NAV format) for use by the agent device.
[0162]
The main window shown in FIG. 28 is an example of a scenario being created, and the initial screen where the scenario editor (240-1-1) is started displays only the start point 308 on the scene development screen 305. Nothing is displayed on the scene screen 301, and the setting screen 303 is not set (default value is displayed).
[0163]
FIG. 29 shows the state of the main window when the branch scene is activated on the scene development screen 305.
As shown in FIG. 29, when a branch scene (branch scene 4 in the figure) is selected and activated, an additional condition classification display column 335 and an additional condition item column 337 are displayed in the area of the scene screen 301 (see FIG. 28). Is displayed.
In the branch condition classification display field 33, the classification of the additional condition item selected for the branch scene in the active state is displayed.
In the branch condition item column, an additional condition item for specifying a branch destination is displayed.
[0164]
FIG. 30 shows the flow of screen operations for editing scenario properties.
In the main window shown in FIG. 28, when the start point 308 displayed on the scene development screen 305 is double-clicked, the scenario property editing window shown in FIG. 30 is displayed over the main window.
[0165]
In this scenario property edit window, scenario name input, kana name input, icon selection, genre selection, priority setting, expiration date (upper limit of time lag from start condition to actual start) setting, running You can set medium execution conditions, scenario start conditions (in a separate window), standby process usage conditions, creator name input, and comment input. The scenario name input and kana name input entered on this screen are management data in the scenario data in the actual machine format.
When the decision button 402 is clicked in the scenario property editing window, the edited contents are reflected in the data, and the process returns to the main window. On the other hand, when the cancel button 403 is clicked, it is not reflected in the data and returns to the main window.
When the standby process use condition setting button 409 is clicked in the scenario property editing window (FIG. 30), a standby process condition screen is displayed, and it is possible to select whether or not to perform the standby process.
[0166]
When a start condition setting button 401 is selected in the scenario property edit window, a main edit window for a scenario start condition (automatic start condition) is displayed (not shown).
The main editing window of the scenario start condition can be set so that the user can start the scenario manually. In this case, uncheck the check box and set it not to start manually.
The automatic start condition (autonomous start condition) list in the main editing window of the scenario start condition displays a condition for the system to automatically start the scenario.
When the new creation button is clicked in the main editing window of the scenario start condition, an automatic start condition selection window (not shown) is displayed and a new start condition can be edited.
[0167]
In the automatic start condition selection window, a judgment condition item (category) to be set is selected, and when the decision is clicked, the process proceeds to a condition range selection window for automatic start. For example, if you want to start a scenario automatically while driving on an expressway (autonomous activation), select the “Choose type” item in “Select when the road condition starts”, and then select “ Select "Highway".
[0168]
Next, various operations for creating a scenario other than the autonomous activation condition will be described.
FIG. 31 shows a flow of screen operations for selecting a screen configuration to be displayed on the agent display screen 51 (see FIG. 14).
When the scene icon 307 displayed on the scene development screen 305 of the main window shown in FIG. 31A is selected and activated, a scene screen 310 corresponding to the selected scene icon is displayed. When the screen configuration change button 309 on the setting screen 303 is clicked, a screen configuration selection window (b) is displayed.
In this screen configuration selection window (b), a list of screen configurations that can be displayed on the scene display screen 54 (see FIG. 18) is displayed. A basic screen with nothing displayed, a 2-choice screen with two selection buttons by buttons, a button-choice screen with multiple selection buttons, for example, a list-selection screen that lists multiple items such as prefecture names Various selectable screens such as an image display screen for displaying image data are displayed.
Select one screen configuration from the listed screen configurations, and click the OK button to change the screen configuration in the confirmation dialog. If you want to change after confirmation, change to that screen configuration and change it to the main window ( Return to a). When returning to the main window, the scene screen 301 is changed to the newly selected screen configuration and displayed.
[0169]
The processes and operations based on FIGS. 32 to 35 form character display contents (images, sounds), screen element creation means for setting screen elements based on the processing contents, and character setting means.
FIG. 32 shows a flow of a screen operation for editing a character action (agent action) instruction.
When the agent display screen 311 is double-clicked with the mouse in the main window (FIG. 28) showing the editing state of the scene screen, the action instruction edit dialog (individual instruction) (FIG. 32 (a)) or the character action instruction edit dialog ( (Unification instruction) (FIG. 32B) is displayed.
Which window is displayed is the window used last time, and when the previous operation instruction is instructed by direct instruction for each character (FIG. 32A) is displayed and expressed in the previous character. When an instruction is given in a state that the user wants to make it, (FIG. 32B) is displayed. When used for the first time, a character action instruction edit dialog (unified instruction) is displayed.
[0170]
In the character action instruction edit dialog (individual instruction) in FIG. 32A, motion (action), facial expression (element of emotion expression), hairstyle (element of growth expression), clothes (element of TPO expression), scale (character display) If the area is a camera frame, the camera angle element), the range to be spoken (the range to which the dialogue is assigned), the operation instruction timing, and the background of the character display area are selected.
When a motion is selected from the motion list, it is displayed in the selected motion column on the right side. When the scene being created is executed, an individual motion image corresponding to the selected motion is reproduced. A plurality of motions can be selected, and individual motion images corresponding to the selected order (list display order of the selected motion column) are reproduced.
In the character action instruction editing dialog (individual instruction) in FIG. 32A, when the determination button is clicked, the edited content is reflected in the data and the process returns to the main window (FIG. 28). Clicking the Cancel button returns to the main window without being reflected in the data.
When the expression content designation button is clicked, the character action instruction edit dialog (unification instruction) is switched to FIG. 32B.
[0171]
When an action instruction (display state) is selected in the character action instruction edit dialog (individual instruction), a scene is defined as a character-specific action. In this case, the agent device 1 determines that it is not a unified action instruction that does not depend on the character in the character drawing / voice output processing by the drawing / voice output unit (101-5).
[0172]
In the character action instruction editing dialog (unified instruction) in FIG. 32 (b), as a state expressed by the character, a unified action instruction table (FIG. 32) that is independent of the characters stored in the common definition DB (240-2-1). 26, see FIG. 27), work element, mental state element, TPO expression element, growth expression element, scale element (when the character display area is a camera frame, the camera angle element) Displayed as selectable. In addition, a screen for selecting the operation instruction timing and the background of the character display area is displayed.
By selecting each display state displayed in the character action state instruction editing window, the user can select a scene in which the display state number corresponding to the display state selected as the action common to each character is set regardless of the character. Is set as the contents of
If the OK button is clicked in this window, the edited contents are reflected in the data and the process returns to the main window (a). Clicking the cancel button returns to the main window (a) without being reflected in the data.
When the direct instruction designation button is clicked, the character action instruction editing dialog (individual instruction) (b) is switched.
[0173]
In the character action instruction editing dialog shown in FIG. 32, a transition mode (quality mode, quick response mode) when an interruption event occurs can be defined in the character action quality column. By defining the transition mode by this character action instruction edit dialog, action interruption setting means for setting whether or not to interrupt the action expression of the character is formed.
When the pull-down button on the left side of the character motion quality column is clicked, selectable modes are “automatic selection”, “switch after waiting for motion” (corresponding to quality mode), “switch without waiting for motion end” ( (Corresponding to the quick response mode), and when one of them is selected, the transition mode corresponding to the individual motion image of the motion selected from the motion list is defined.
Note that by specifying the beginning (left side) and end (right side) of the line assignment range in the character action instruction editing dialog (individual instruction) in FIG. 32A, the individual action image corresponding to the motion in the specified range. Is associated with the dialogue (character voice) set in the scene being edited. As a result, the individual motion images in the designated range are reproduced by matching the associated dialogue.
For example, when five motions are selected from the motion list in the order of a, b, c, d, and e, and 2 to 4 are specified as ranges as the line assignment, the reproduction of the individual motion image a is finished. Later, the individual motion images from b to d are reproduced, and during this time, the character outputs speech as speech. Thereafter, the individual operation image e is reproduced.
[0174]
FIG. 33 is an explanatory diagram of the voice editing memo window shown when the dialogue editing button is selected in the main window.
When the dialogue editing button 313 is selected (clicked) in the main window of FIG. 28, the scenario editor 211 newly displays a voice editing memo window 600 shown in FIG.
Although described as the process of the scenario editor 211, what is actually displayed is the scenario editor 211 (program) and the CPU 101 cooperating to perform processing such as display, but in order to simplify the explanation, the operations of the two cooperated Will be described as processing and operation of the scenario editor 211. The following description and the process and operation of the synthesized speech editing dialog 214 will be described in the same manner.
[0175]
When the voice editing button 601 is selected in the voice editing memo window 600 of FIG. 33, the synthesized voice editing dialog 214 is activated, and the synthesized voice editing dialog 214 displays the voice editing main screen 608 shown in FIG.
When synthesized voice data is created on the voice editing main screen 608, the scenario editor 211 displays a text corresponding to the synthesized voice data on the balloon display portion 602 of the voice editing memo window 600.
It is also possible to directly enter text in the balloon display unit 602 without performing voice editing. In this case, since the synthesized voice data is not created, even if this scenario is executed by the agent device, no voice is output, and the text input from the balloon display unit 602 is displayed in the balloon 52.
[0176]
FIG. 34 shows the audio editing main screen 608.
The voice editing main screen 508 includes a voice specifying unit, a dialogue input unit, a front matching search unit, a result display unit, and a cancel button.
The voice specifying unit includes a voice type selection button 607 for selecting a voice type and an emotion selection button 605 for selecting an emotion.
The dialogue input unit includes a dialogue input box 610 and a conversion button 611.
The forward match search unit includes a check box 612, a candidate list box 613 in which forward match candidates are displayed as a list, a candidate number box 614 in which the number of all candidates including a candidate not listed is displayed, and candidates displayed in a list It is comprised by the candidate selection button 615 which selects.
The result display unit is a word string information box 619 that visually displays the connection of sounds by the synthesized voice data by a word surrounded by a frame and a silent line segment, a play button 620, a delete all button 621, a sentence registration button 622, A registered text list button 623, a selected word registration button 624, a new word registration button 625, a registered word list button 626, and an enter button 627 are configured.
[0177]
By inputting a speech that makes up the content of the conversation that the agent wants to pronounce in the speech input box 610 of the speech editing main screen 605, the corresponding synthesized speech data is created and edited. Hereinafter, the synthetic voice creation and editing processing operations will be described.
[0178]
When the voice type selection button 607 is selected, the synthesized voice editing dialog 214 displays a drop-down list of selectable voice types. As a selectable voice type, one of adult female, adult male,..., Automatic selection can be selected, and the selected voice type is stored in the memory 102 as synthesized voice data. The default sound type is defined for adult women.
When automatic selection is selected as the type, when the synthesized voice data created and edited in the synthesized voice editing dialog 214 is auditioned (when the playback button 620 is clicked), it is played back with the default type.
The agent device that executes the scenario also has a function to specify the voice. When free selection is selected, it conforms to the voice specified on the agent side, the voice specified by the agent device user, or the specified character. Voice (voice type) is output. On the other hand, when a voice type other than free selection is selected as the synthesized voice data of the scenario data, the voice type voice according to the synthesized voice data is preferentially output regardless of the designation by the agent device. .
[0179]
When the emotion selection button 605 is selected, the synthesized speech editing dialog 214 displays a drop-down menu of emotions indicating emotions such as fun, sad, energetic,.
When any emotion is selected by the user, the synthesized speech editing dialog 214 preferentially selects the speech code corresponding to the selected emotion from the unit speech data that exists for the same notation from the dialogue DB 221a. To search.
If no emotion is selected, the emotion “normal” is selected.
In this way, the designation means of the present invention is formed by designating at least one of emotion and voice type.
[0180]
When the dialogue input box 610 is designated and characters are input, the synthesized speech editing dialog 214 sequentially displays the input characters.
When the check box 612 is checked, the notation character portion of the dialogue DB 221a is searched for in front, the candidate notation is displayed in a list in the candidate list box 613, and the number of candidates listed is displayed in the candidate number box 614. indicate.
In the example displayed in FIG. 34, since “Today” is input in the dialogue input box 610, the synthesized speech editing dialog 214 displays “Today”, “Today”, and “Today” as candidates that “Today” matches forward. 16 candidates such as “is fine” are searched, and 8 of them are displayed. Candidates that are not displayed can be scrolled by moving the scroll button on the right side of the candidate list box 613.
For example, when “Today is sunny” is displayed in the dialogue input box 610 to which “has fine” is further input, as a result of the forward matching search, “Today is sunny”, “Today is sunny” and “ Three candidates, “It ’s sunny today,” are listed.
As described above, the forward matching search is automatically executed without any user operation in accordance with the text input and displayed in the dialogue input box 610.
[0181]
When any one of the forward matching candidate lists displayed in the candidate list box 613 is selected and determined, the word list information box 619 displays the dialogue input box 610, the candidate list box 613, and the candidate number box. The display of 614 is cleared.
The selected candidate is determined by mouse double-clicking the candidate, selecting a candidate, clicking the candidate selection button 615, selecting a candidate, and then selecting the candidate using the “Enter” key on the keyboard.
[0182]
Next, creation of synthesized speech data and word string display for visually displaying an input sentence displayed in the word string information box 619 will be described.
When the conversion button 611 is selected, the synthesized speech editing dialog 214 converts the text (input sentence) displayed in the dialogue input box 610 as a conversion target sentence in a batch.
If the text sentence to be displayed on the balloon screen 52 (see FIGS. 14 and 28) has already been input (the synthesized speech data may not be created), the dialogue edit button 313 in FIG. When clicked, the input text sentence is displayed on the balloon display unit 602. When the voice editing button 601 is selected in this state, the synthesized voice editing dialog 214 displays the text sentence displayed in the balloon display unit 602 in the dialogue input box 610. As a result, the synthesized voice data that has already been created can also be edited.
[0183]
FIG. 35 shows the flow of screen operations for editing the speech recognition dictionary.
This operation is to set a voice dictionary for recognizing a voice response returned from the user by the agent device when the agent device asks for a response based on the created scenario.
In the main window showing the editing state of the scene screen (FIG. 28), when a button part portion 315a (which may be a normal list box part portion depending on the screen configuration) displayed according to the selected screen configuration is double-clicked, the sound is displayed. A recognition dictionary selection window (FIG. 35A) is displayed. The speech recognition dictionary selection window is also displayed by double-clicking the dictionary list display section 315b recognized in the background.
[0184]
In the speech recognition dictionary selection window (FIG. 35 (a)), when a dictionary name in the dictionary candidate list display is double-clicked, it is displayed in the list selected as a general dictionary as using the speech recognition dictionary.
When the OK button is clicked, the edited content is reflected in the data and the screen returns to the main window (FIG. 28), and when the Cancel button is clicked, the data is not reflected in the data and the screen returns to the main window.
When the user-defined dictionary edit button is clicked, a voice recognition dictionary creation window (FIG. 35B) for creating a new voice recognition dictionary is displayed. In this window, when a dictionary name is input and the dictionary addition button is clicked, a window for creating a new speech recognition dictionary with the name and registering the word in the speech recognition dictionary (FIG. 35C) is displayed.
When the OK button is clicked in the speech recognition dictionary creation window, creation of the speech recognition dictionary is terminated, and the speech recognition dictionary selection window returns.
[0185]
In the window for registering words in the speech recognition dictionary (FIG. 35 (c)), the word to be registered is input into the reading field in half-width kana and the decision button is clicked. Next, a name (name to be displayed) is selected or newly entered, and a PCM voice for callback is selected (if none is selected, TTS is used for callback). After entering these three items, clicking the registration button will register the data and add it to the registered word list on the right.
When you have finished registering the words you want to register, click the Back button to return to the voice recognition dictionary creation window.
[0186]
Next, an operation for editing a scenario flow will be described.
In the main window shown in FIG. 28, the scene icon 307 being created is selected and activated. When the new scene creation button 323 is clicked in this state, a transition selection window (not shown) is displayed.
In this transition selection window, the condition for branching to the newly created scene is selected from the branch event list to determine the transition condition to the next scene (newly created scene), and the process returns to the main window.
A new scene is created on the scene development screen 305 of the main window after returning, and is displayed as NEW to distinguish it from other scene icons.
The branch events that can be selected in the branch event selection window are displayed in FIG.
As described above, by setting the conditions for shifting from one screen element to the next screen element, the transition condition setting means in the present invention and the time limit for shifting from one motion process of the character to the next motion process The transition time limit setting means for setting is formed.
[0187]
Next, an operation for editing the end position of the scenario will be described.
When the scenario end ID button 327 is clicked in the main window of FIG. 28, an end ID designation window (not shown) is displayed.
In this end ID designation window, the ID number of the end position mark is designated. Normally, it is automatically assigned, but the editor operator can also assign it by unchecking the check box written as automatic assignment. When the OK button is clicked, the ID number is determined and a branch event selection window (not shown) is displayed.
In the branch event selection window, the branch condition for ending the scenario is set in the same way as when creating a new scene. Additional conditions can be set in the same way. If an OK button is clicked in this window, the condition (transition condition) is determined and the process returns to the main window (transition condition setting means). At this time, a new end ID is newly created and displayed on the scene development screen 305 of the main window.
As described above, a screen element in which at least one of display contents and processing contents of a character (agent) is defined is defined as one screen element (scene), and the screen element is combined with the transition condition between the screen elements and the screen element. A screen element transition body creating means for creating a transition body (scenario) is formed.
Further, as described above, the character display processing setting means for setting the processing content of the character to be displayed on the display device in the vehicle is formed.
[0188]
FIG. 36 shows a flow of a screen operation for compiling the created scenario into a real machine format (NAV format) that can be used by the agent device.
When the build button 332 is clicked in the main window (FIG. 36A), a scenario compiler window (b) is displayed.
In this scenario compiler window (b), specify the name of the file that outputs the compiled data, select the scenario to be converted at the same time (convert the scenario checked in the scenario list), and click the compile button. The scenario compiler (240-1-2) starts data conversion. The data conversion status is displayed on the result display section.
When the end button is clicked, the data conversion is completed and the process returns to the main window (a).
[0189]
As described above, the screen element transition body creating means for creating the screen element transition body (scenario) by combining the screen element (scene), the transition condition, the branch element, and the branch condition is formed.
Further, as described above, the character display processing setting means for setting the processing content of the character to be displayed on the display device in the vehicle is formed.
[0190]
As described above, according to the scenario creation device of the present embodiment, the display state instructing the action of the character in each scene of the scenario is made common regardless of the type of the character, and thus can be executed without being bound by the character. Scenarios can be created, and scenarios created for each character can be combined into one, making scenario creation easier.
[0191]
In addition, according to the agent device of the embodiment described above, the long-term emotional element and the short-term emotional element are defined as the psychological state of the agent, and both the emotional elements are referred to, and the action is determined to make the agent more human Can behave.
In the scenario creation device, both emotion elements can be set as scenario transition conditions, and the scenario of an agent that behaves more like a human can be created by making both emotion conditions changeable in the scenario.
[0192]
In addition, according to the present embodiment, the process for determining whether the condition for autonomously starting (automatically appearing) the agent based on the scenario data created by the scenario creating device is satisfied periodically or in a specific state It can be executed when the condition is satisfied, and the agent can be made to appear automatically when the condition is satisfied.
On the other hand, according to the scenario creation device and the scenario editor of this embodiment, regardless of whether or not there is knowledge of the program, having a scenario editor makes it possible for agents to automatically appear and respond when a specific condition is met. Scenario data can be easily created and edited.
[0193]
Further, according to the present embodiment, since each individual motion image starts with the basic posture state image and ends with the basic posture state image, the posture of the character does not change abruptly between the individual motion images. , Can be more natural movement.
In addition, it is possible to adjust the time without performing an unnatural operation by configuring the individual motion images as a start moving image, a retained moving image, and an end moving image, and repeating the reproduction of the retained moving image.
If the response speed is more important than quality, the character's response can be made quicker by interrupting the individual motion image being reproduced and reproducing the next individual motion image when the interruption condition is satisfied. .
[0194]
Although the preferred embodiments of the present invention have been described above, various modifications can be made in the present invention.
For example, the reproduction after interruption in the quality mode may be the second quality mode II as shown in FIG.
Note that the second quality mode II may be adopted instead of the quality mode described with reference to FIG. 20 of the embodiment, and the quality mode of FIG. 20 and the second quality mode II may be used in combination. When used together, the second quality mode II is added to the selection object in the character action quality pull-down menu of FIG. Further, in the agent device, “quality-oriented” (corresponding to the quality mode in FIG. 20) and “reaction-oriented” (as shown in FIG. 13) are automatically selected for the operation switching determination data (10-2-3-8) (see FIG. 13). This is defined by adding “quality / intermediate reaction” (corresponding to the second quality mode II) to the quick response mode of FIG.
[0195]
FIG. 37 shows the reproduction state after interruption in the second quality mode II.
(1) When starting video playback
As shown in FIG. 37A, when an interruption event occurs during the reproduction of the start moving image, the starting moving image being reproduced is interrupted and the end moving image is reproduced from the middle.
The playback start position of the end video is the playback time of the playback video and the end video, and when the time from the playback of the start video to the occurrence of the interruption event is T2, the end video is after the time T1-T2 has elapsed. Start from the point. In this case, the playback time of the end moving image is also T2.
The start video and the end video are in a symmetrical relationship, that is, when the end video is played in the reverse direction from the basic posture state image, it becomes the same as the start video, so the image when the start video is interrupted and the end video are started It matches the image at the time. For this reason, the gap of the image at the time of interruption can be eliminated, and the image being reproduced can be ended early.
[0196]
(2) During playback of retained video
As shown in FIG. 37 (b), when an interruption event occurs during the reproduction of the retained moving image, the retained moving image being reproduced is interrupted and the end movie is reproduced, and then the process proceeds to the subsequent individual operation image. To do.
As a result, the remaining playback time of the retained moving image can be promptly adapted. In this case, since the retained moving image is slightly moved from the retained state image and the end moving image starts from the retained state image, the gap between the images is reduced. As a slight operation from the holding state images 12b1 and 12b4 in this case, the movement is within a range where the difference from the holding state images does not exceed 10%.
As a slight movement within the predetermined range, the difference from the previous image in unit time may be within 10%. Further, the difference range may be 5%, 15%, or 20% in addition to 10%.
(3) End video playback
As shown in FIG. 37 (c), when an interruption event occurs during the reproduction of the end moving image, the end moving image is reproduced as it is, and then the subsequent individual operation image is transferred. This operation is the same as the transition shown in FIG.
In the retained video for time adjustment in the case of the second quality mode II described above, the agent is operating because it performs a slight operation such as waving a small hand, blinking, squeezing the small neck, etc. This can be recognized by the user. Further, since the retained moving image has a slight movement in the retained state, the gap between the images does not increase even if the retained moving image is interrupted in the middle of repeated reproduction and is shifted to the end moving image.
[0197]
Moreover, although embodiment demonstrated demonstrated the case where there was one basic attitude state, you may make it provide two or more basic attitude states.
For example, when there are three basic posture states A, B, and C, individual motion images that start from the respective basic posture states A, B, and C, return to the basic posture states A, B, and C through the holding state Prepare. Each individual motion image is composed of a start moving image, a retained moving image, and an end moving image, as in the embodiment described above.
In addition, a basic posture change moving image for changing from one basic posture state to another basic posture state is separately prepared. If the number of basic posture states is Q, (Q ² -Q) Required. For example, when there are three basic posture states A, B, and C, a basic posture change video that starts at A and ends at B, a basic posture change video that starts at B and ends at A, and starts at B and ends at C Basic posture change video that ends, basic posture change video that starts at C and ends at B, basic posture change video that starts at C and ends at A, and basic posture change video that starts at A and ends at C Prepare a posture change video.
Then, it is possible to continuously reproduce a moving image with no gap between images using the basic posture changing moving image and the individual operation image.
By providing a plurality of basic posture states in this way, it becomes possible to cause a character to perform a complicated motion.
[0198]
【The invention's effect】
According to the invention described in claims 1 to 3, one screen element of the screen element transition body Action expression video defined in If there is a predetermined instruction from the in-vehicle device or user during execution of Action expression Subsequent moving images can be reproduced with less unnaturalness with respect to the reproduction quality and response speed of the moving images.
According to invention of Claim 4, Claim 5, while creating the screen element transition body which a character communicates by being performed by a vehicle-mounted apparatus, and its starting condition, 1 Screen element Action expression video defined in Running ,car Loading device Or its user Character's motion expression when given instructions from Video Whether to interrupt About mode Can be set.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of an agent device according to an embodiment of the present invention.
FIG. 2 is a configuration diagram of various status detection devices in the agent device.
FIG. 3 is an explanatory diagram illustrating a relationship between an agent processing unit realized by a program being executed by a CPU and an overall processing unit.
FIG. 4 is an explanatory diagram showing a configuration of an agent processing unit.
FIG. 5 is an explanatory diagram conceptually showing information collected in an external storage medium.
FIG. 6 is an explanatory diagram showing the contents of mental model data in a conceptual manner.
FIG. 7 is an explanatory diagram conceptually showing long-term emotional elements.
FIG. 8 is an explanatory diagram exemplifying the contents of regulation of long-term emotion change conditions.
FIG. 9 is an explanatory diagram showing short-term emotional elements as providing concepts.
FIG. 10 is an explanatory view exemplifying the contents of provisions for short-term emotion change conditions.
FIG. 11 shows a configuration of actual machine format scenario data.
FIG. 12 is an explanatory diagram showing an example of the content of an individual action image representing an action in which a character raises his right hand.
FIG. 13 is an explanatory diagram conceptually showing the contents of operation switching determination data in agent data.
FIG. 14 is an explanatory diagram showing an example of a scene screen displayed on a display device based on scenario scene data.
FIG. 15 is a screen transition diagram showing scene screen transitions for each scene according to a guidance scenario transmitted by the inn to a prospective guest.
FIG. 16 is a flowchart showing an example of a flow of scenario execution processing.
FIG. 17 is a flowchart showing an example of the flow of scene processing.
FIG. 18 is a flowchart showing character drawing / sound output processing by a drawing / sound output unit (101-5).
FIG. 19 is an explanatory diagram showing a flow that can be transferred between a start moving image, a retained moving image, and an end moving image when an individual operation image is reproduced.
FIG. 20 is an explanatory diagram showing a reproduction state after interruption in the quality mode.
FIG. 21 is an explanatory diagram showing a reproduction state after interruption in the quick response mode.
FIG. 22 is a configuration diagram of a scenario creation device.
FIG. 23 conceptually shows the configuration of a scenario editing program and data.
FIG. 24 is an explanatory diagram conceptually showing conversion of a data format.
FIG. 25 is a development condition item table storing development condition items (transition conditions) for branching from one scene to the next scene (scene development).
FIG. 26 is an explanatory diagram conceptually showing a part of the contents of a character display state instruction table stored in a common definition DB.
FIG. 27 is an explanatory diagram conceptually showing another part of the contents of the character display state instruction table stored in the common definition DB.
FIG. 28 shows a configuration of a main window displayed on a display device when a scenario editor is activated.
FIG. 29 is an explanatory diagram showing the state of the main window when a branch scene on the scene development screen is activated.
FIG. 30 shows a flow of a screen operation for editing a scenario property.
FIG. 31 shows a flow of a screen operation for selecting a screen configuration to be displayed on the agent display screen.
FIG. 32 shows a flow of a screen operation for editing a character action (agent action) instruction.
FIG. 33 is an explanatory diagram of a voice editing memo window shown when a dialogue editing button is selected in the main window.
FIG. 34 is an explanatory diagram of a voice editing main screen.
FIG. 35 shows the flow of screen operations for editing a speech recognition dictionary.
FIG. 36 shows a flow of screen operations for compiling a created scenario into a format of an actual machine format that can be used for navigation.
FIG. 37 is an explanatory diagram showing a modified example of the reproduction state after interruption in the quality mode.
[Explanation of symbols]
1 Agent device
2 Scenario creation device
3 servers
(1) Central processing unit
(2) Display device
(3) Audio output device
(4) Voice input device
(5) Input device
(6) Various situation detection devices
(7) Various in-vehicle devices
(8) Communication control device
(9) Communication device
(10) External storage device
(200) Control unit
(210) Input device
(220) Output device
(230) Communication control device
(240) Storage device
(250) Storage medium drive
(260) Input / output I / F

Claims (5)

Start moving from the basic posture of the character to a predetermined posture state indicating a representation content, the predetermined posture state holding video for holding at least the operation of the character of three video ends moving from the predetermined posture to the basic posture state an operation expression moving image storage means for storing a representation videos,
Screen element transition body storage means for storing a screen element transition body configured by combining the screen elements with a screen element in which display content including a motion expression moving image of the character is defined as one screen element;
Screen element transition body executing means for executing the screen element transition body;
During execution of the defined operating expressed videos on one screen element of the screen element transition body, when a predetermined instruction from the vehicle device or the user, the next screen element to interrupt the operation expressions video being the execution a readiness mode in which the execution and the operation representation moving in the execution has been completed, the mode determining means for determining a quality mode for executing the next screen element, and
A vehicle-mounted device comprising: The mode determination means determines whether the mode is a quick response mode or a quality mode from a mode defined by one screen element of a motion expression video being executed when a predetermined instruction is received from an in-vehicle device or a user. The in-vehicle device according to claim 1. The mode determining means determines whether the mode is a quick response mode or a quality mode according to a table defined in the in- vehicle device when there is a predetermined instruction from the in-vehicle device or the user.
The in-vehicle device according to claim 1. A data creation device for creating a screen element transition body executed by an in-vehicle device,
Specified in a stage comprising at least three moving images: a starting moving image from the basic posture state to a predetermined posture state indicating the contents of expression, a holding moving image that holds the predetermined posture state, and an ending moving image from the predetermined posture state to the basic posture state. A screen element creating means for creating a screen element in which display contents including a motion expression video of a character are defined as one screen element ;
Transition condition setting means for setting a transition condition for transitioning from one screen element created by the screen element creation means to the next one screen element;
Screen element transition body creating means for creating a screen element transition body to be executed by the in-vehicle device based on the one screen element and the transition condition;
It said screen element creation means, executing a defined operation expressed moving by the one screen element, when a predetermined instruction from the vehicle-mounted device or the user has been made, whether or not to interrupt the operation representation video Character A data creation apparatus, further comprising an operation interruption setting means for setting the mode of 1 to the one screen element . The operation interruption setting means is an immediate mode for interrupting an action expression movie being executed and executing the next screen element, a quality mode for executing the next screen element after continuing the action expression movie being executed, whether to interrupt Set any one of the automatic determination modes to determine whether or not on- vehicle device side,
The data creation device according to claim 4.

Images