JP4984708B2 - Information processing apparatus having voice dialogue function - Google Patents

Description

  The present invention relates to an information processing apparatus or electronic device having a voice interaction function, and in particular, in a car navigation system and an information processing apparatus having a voice service function, a low priority voice according to a priority of voice processing of a plurality of services. The present invention relates to a technique for executing audio processing with different priorities at an appropriate timing without interrupting the processing.

  Car navigation systems and telematics terminals for vehicles not only provide unilateral information to the user like a radio, but also process input operations such as voice input and button input from the user in an interactive format. Information can be provided. Thereby, the user can use various information providing services such as route guidance, weather forecast, and news. Further, even when a plurality of services are provided, the user can use the services to be switched while specifying the services by following the hierarchy of the service menu or by directly calling the services.

  In particular, in the use of services in the vehicle, the use of voice interfaces is emphasized from the viewpoint of not hindering driving. For this reason, a car navigation system and a telematics terminal capable of interactively using voice to provide information to users and service operations have been already provided.

  In an ordinary interactive information processing system, various services are provided to a user by processing a service scenario in which an operation of the system for providing a user's desired service is described. In general, a service scenario is expressed by a state transition model, service output information is assigned to each state, and user input information candidates are assigned to each transition. The interactive information processing system outputs the contents of the output information of the service assigned to the current state in the state transition model, and transitions to the next state according to the user input information input in response. Then, the dialogue is advanced by repeating this.

  The service scenario is described by a voice interaction description language such as VoiceXML. Candidates for user input information items are described in a grammar description language such as SRGS (Speech Recognition Grammar Specification). The VoiceXML interpreter interprets a service scenario and interactively executes information provision, reservation, device operation, and the like according to information input by the user.

  An interactive information processing system can handle various types of user input. For example, voice recognition technology is used to accept a user's voice input. In the speech recognition technology, a speech recognition module (ASR, Automatic Speech Recognition) converts user input speech input during a predetermined input period into text information. The speech recognition module refers to the designated input item candidate, and when it is estimated that the user input matches any of the input candidates, the speech recognition module regards it as a recognition result. In addition, a voice synthesis technique is used to output a voice to the user from the system. In the speech synthesis technology, text information is automatically converted into output speech and output in a predetermined output period by a speech synthesis module (TTS, Text To Speech). The user uses not only input devices and output devices directly coupled to the interactive system, but also mobile communication devices such as fixed phones, mobile phones, PDAs, and information processing devices such as car navigation devices, An information processing service can be used by connecting to an interactive system via a communication network.

Japanese Unexamined Patent Publication No. 10-47984 (Patent Document 1) describes a car navigation device. The route guidance timing control device of the car navigation device calculates the time interval Tw, a register for setting the reference time Ts, a register for setting the time interval Tw for calculating the arrival time Tr to the intersection where the vehicle should turn, and the time interval Tw. And a register for setting the minimum time unit Ti. The route guidance of the car navigation device is calculated using the running state information of the own vehicle, the position information of the own vehicle, and the position information of the intersection to be guided as judgment elements, and calculates the time for the own vehicle to reach the intersection to be guided. The route guidance is transmitted to the driver by changing the timing of outputting the route guidance. Thereby, the reliability of route guidance corresponding to the type of road at the intersection to bend and the congestion state of the road is improved. This method works well when a single service performs only a single voice notification, but does not take into account the case where a plurality of voice notifications are generated using a plurality of services.
Japanese Patent Laid-Open No. 10-47984

Japanese Patent Laying-Open No. 2000-304553 (Patent Document 2) describes a roadside communication device. When a service request instruction is input from a navigation device as an in-vehicle device, the service request is output toward the roadside. When there is a delay instruction response from the information center as a road device, the navigation device responds within the delay time. Search for and output other services that will save time, and receive the results. A response to a service request made earlier after the delay time has elapsed is received. Therefore, even when there is a service that requires time for processing on the roadside device side, other services and information can be exchanged within the delay time. As a result, information is exchanged more quickly and efficiently between the road side and the moving body side. With this method, it is possible to use a plurality of services in parallel, but each time a process that takes time occurs in the system, another service operates, so service interruptions frequently occur.
JP 2000-304553 A

Japanese Patent Laying-Open No. 2003-202233 (Patent Document 3) describes an information reproduction method that is mounted on a vehicle and performs speech synthesis based on the content of a text sentence. The information reproduction method acquires state information of the vehicle speed and the vehicle position, calculates a time limit based on the state information, and text that can be uttered within the time limit of the text sentence to be uttered. A sentence is selected from a predetermined storage means that stores a plurality of text sentences with utterance contents, and a speech synthesis parameter is generated from the selected text sentence and output as a voice. In this method, since an output that falls within the time limit is selected, one audio output is not interrupted. However, the dialogue process does not necessarily end with only one notification, and may be configured by multiple responses with the user. This method does not guarantee that such a batch of dialogs is executed without interruption.
JP 2003-202233 A

  In a car navigation system or a telematics terminal for vehicles, when multiple services are used in parallel while switching, or when multiple services automatically give voice notifications as necessary, during the audio output of multiple services May cause contention. That is, while a user is using a certain voice service, both services may take away the right of voice output when another service tries to make a voice notification. For example, while the user is searching for a restaurant while driving, the route guidance service may interrupt and notify the voice guidance of an intersection to turn right.

  In such a case, immediately after the restaurant search service asks the user a question for narrowing down the search target for creating the search condition, the right / left turn guidance is notified, and the same question is given again and voice input is made to the user again. Dialogue proceeds in the order of requesting. Therefore, when the user tries to input the answer to the question by voice, the voice notification of the left / right turn guidance prevents the user from thinking and inputting the answer, so that the user cannot have a smooth dialogue with the restaurant search service. If the interrupted voice notification can be delayed, the problem may be solved. However, if the timing of high-priority right or left turn guidance is delayed, it will not be useful as route guidance. Therefore, it is necessary to give priority to voice guidance with high priority.

  If the inventors determine the timing of each voice processing according to the priority of voice processing (voice output and / or voice input) of a plurality of services and each time length required for the plurality of voice processing, Voice processing was performed at the timing, and the user recognized that multiple services could be used smoothly.

An object of the present invention is to prevent conflicts between the plurality of audio processing.

Another object of the present invention is smoothly that multiple voice processing at each proper timing for the user.

According to the features of the present invention, an information processing apparatus having a voice interaction function includes service scenario storage means for storing a plurality of service scenarios having classifications and priorities, and high priority in the service scenario storage means. Notification time estimating means for estimating the timing of the next first voice processing according to the service scenario, free time estimating means for estimating the free time from the present time to the timing of the first voice processing, and storing the service scenario Dialog time estimation means for estimating the time length required for the second voice processing including voice output and processing for accepting user input according to a low priority service scenario in the means, and the estimation of the second voice processing If the estimated duration is less than the estimated free time, the free time Dialogue control means for performing the second speech processing according-bis scenario, which comprises a.

According to the present invention, it is possible to prevent contention between a plurality of audio processing, it is smoothly that multiple voice processing at each proper timing for the user.

  Embodiments of the present invention will be described with reference to the drawings. In the drawings, similar components are given the same reference numerals.

  FIG. 1 shows a configuration of an information processing apparatus or electronic device 10 having a voice interaction system 12 that performs voice processing, that is, voice output and voice input, of a plurality of voice services 400 according to the present invention. The block diagram of FIG. 1 can also be viewed as a flow diagram for voice interaction processing performed by the processor 100. The information processing apparatus or the electronic device 10 may be, for example, a car navigation apparatus, an information processing apparatus having a car navigation function, or a telematics terminal. The information processing apparatus or electronic device 10 includes a processor 100, a memory 102, a display 104, an acoustic device 106 including a speaker and a microphone, and an input device 108 including keys and buttons. Typically, a voice interaction system 12 is implemented on the processor 100 and the memory 102. The information processing apparatus or the electronic device 10 implements service functions such as car navigation, vehicle maintenance, restaurant search, facility search, e-mail, news, audio playback, and the like. These service functions include a respective voice service 400.

  The voice dialogue system 12 is provided for each of a plurality of voice services 400, and a notification time estimation unit 20 that estimates a time and a length of time for voice notification. The voice dialogue system 12 presents a voice notification with a high priority at a certain time. A free time estimation unit 30 that estimates a remaining free time t1 from the voice notification to the voice notification and further estimates a free time t1 between two high priority voice notifications, and a plurality of service scenarios in the plurality of voice services 400. Stores service scenarios including each interaction classification (eg, notification, question requesting answer, question requesting answer and confirmation notice for answer) and priority (eg, 0, 1, 2,...) Service scenario storage unit 40, dialogue time estimation unit for estimating time t2 required for dialogue of a service scenario about to output voice 0, dialogue history information, information storage unit 60 that stores the user response time history data and incoming text, the dialogue control unit 70 which determines the sound processing to be executed, and includes an input processing unit 72 and the output processing section 74. The dialogue time estimation unit 50 includes an input time estimation unit 52 and a character number counter 54. The functions of the components 20, 30, 50, 70, 72, and 74 can be implemented on the processor 100 according to a program stored in the memory 102.

  The notification time estimation unit 20 is connected to the free time estimation unit 30 and the service scenario storage unit 40. The dialogue time estimation unit 50 is connected to the service scenario storage unit 40. The dialogue control unit 70 is connected to the service scenario storage unit 40, the free time estimation unit 30 and the dialogue time estimation unit 50, and is further connected to an input processing unit 72 and an output processing unit 74. The dialogue control unit 70 determines which service scenario of the voice service 400 currently being executed is to be executed based on the idle time t1 from the idle time estimation unit 30 and the dialogue time t2 from the dialogue time estimation unit 50. Then, the input processing unit 72 and the output processing unit 74 are controlled according to the determined service scenario.

  As an example, it is assumed that two voice services A and B are operating in parallel, and the voice processes a and b of the two voice services can compete in time. The notification time estimation unit 20 for each service generates a high priority voice notification a to be uttered next in time for the voice service A based on the service scenarios of the voice services A and B in the service scenario storage unit 40. A predicted time tp to be performed is estimated. The available time estimation unit 30 calculates the current available time t1 = tp−tc by subtracting the current time tc from the predicted time tp. The dialogue time estimation unit 50 calculates a time t2 required for the voice processing b (voice notification and input) with the lower priority to be uttered next in the voice service B. The voice processing b has a low priority and may be executed earlier or later than the predicted time tp. The voice processing b is performed in units of dialogue processing based on dialogue classification (for example, notification, question for requesting an answer, question for requesting an answer, and confirmation notification for the answer) set in the service scenario.

  For example, when the dialogue classification of the service scenario of the voice service B is “notification”, the dialogue processing unit (group) is voice synthesis output processing of the text of the notification. For example, when the dialogue classification of the service scenario of the voice service B is “question that requires an answer”, the dialogue processing unit is a voice synthesis output process of a question text message and a voice recognition of the answer message of the user corresponding thereto. It is processing. For example, when the dialogue classification of the service scenario of the voice service B is “question for requesting an answer and a confirmation notification for the answer”, the dialogue processing unit includes the voice synthesis output processing of the text message of the question and the user's A voice recognition process for the answer message and a voice synthesis output process for the confirmation text message for the answer.

  The time required for uttering the text voice can be calculated based on the number of characters or the number of mora. The time required for the response message can be the user response time history information, general user response time data for a particular question, the predicted user response time preset in the service scenario, or the longest possible response text message Can be calculated based on the number of characters or the number of mora.

  The dialogue control unit 70 sets the idle time t1 until the voice notification a with high priority in the voice service A based on the service scenario from the service scenario storage unit 40 and the voice processing b with low priority in the voice service B. The required audio processing time t2 is compared. If the relation of voice processing time t2 <free time t1 is satisfied, the dialogue control unit 70 immediately performs voice processing b with a low priority of the voice service B using the input unit 12 and / or the output unit 14. On the other hand, if the voice processing time t2 ≧ the idle time t1, the dialogue control unit 70 suppresses the voice processing b with a low priority of the voice service B, and immediately or at the scheduled time, the voice in the voice service A first. Notification a is performed. Thereafter, the voice interaction system 12 repeats the same operation until the voice processing time t2 <the idle time t1 and the voice processing b with a lower priority of the voice service B is executed.

  In parallel, three or more voice services A, B, C,. . . In the same manner, the voice processing a, b, c,. . . Can be executed. Assume that voice processing a of voice service A has the highest priority, voice processing b of voice service B has the next highest priority, and voice processing c of voice service C has the lowest priority. . In this case, the voice processing a of the voice service A is executed with priority.

  When the user uses the voice service B, the dialogue control unit 70 sets the time t2 required for the voice processing b of the voice service B to the free time t1 before the predicted time tp of the first voice processing a of the voice service A. If the voice processing time t2 <the idle time t1, the voice processing b of the voice service B is executed during the idle time t1. When the voice processing time t2 ≧ the idle time t1, the dialogue control unit 70 inhibits (prohibits) the voice processing b of the voice service B and first executes the voice processing a of the voice service A. Thereafter, the procedure is repeated until the voice processing b of the voice service B is executed because the voice processing time t2 <the idle time t1.

  When the user uses the voice service C, the dialogue control unit 70 sets the time t2 required for the voice processing c of the voice service C before the predicted time tp of the first voice processing a or b of the voice services A and B. When the voice processing time t2 <the idle time t1, the voice processing c of the voice service C is executed during the idle time t1. When the voice processing time t2 ≧ the idle time t1, the dialogue control unit 70 suppresses the voice processing c of the voice service C and first executes the first voice processing a or b of the voice service A or B. Thereafter, the procedure is repeated until the voice processing c of the voice service C is executed because the voice processing time t2 <the idle time t1.

  FIG. 2 illustrates an information processing apparatus or electronic device 10 having a voice interaction system 12 that performs voice processing, that is, voice output and input, of a route guidance service 402 and an e-mail service 404 of a car navigation system according to an embodiment of the present invention. The configuration is shown. In this case, high-priority voice notification by the route guidance service 402 is predicted based on the position of the vehicle and the surrounding environment obtained from the map database, web page, or text broadcast traffic information.

  The information processing apparatus or electronic device 10 includes a car navigation system information storage unit 80. The car navigation system information storage unit 80 stores route information 82, a current position 84, and a map database 86. The notification time estimation unit 20 includes a next guidance time estimation unit 22. The next guidance time estimation unit 22 is connected to the car navigation system information storage unit 80 and the vehicle speed sensor 112. Other configurations of the information processing apparatus or the electronic device 10 are the same as those in FIG.

  The route guidance service 402 gives priority to voice notification, such as a traffic jam warning, a speed warning when entering a curve, a warning of a traffic accident frequent section, a warning of a speed control section, a road surface freezing warning, a falling object warning, etc. May be included.

  For the route guidance service 402, the next guidance time estimation unit 22 uses, for example, a passing point in the route (for example, from the vehicle speed detected by the vehicle speed sensor 112 and the route information 82 in the car navigation system information storage unit 80). Estimate the arrival time to the intersection). The next guidance time estimation unit 22 estimates a predicted time tp at which a voice notification event of the route guidance service 402 with a high priority should be generated based on the currently predicted driving situation. The available time estimation unit 30 calculates the current available time t1 = tp−tc by subtracting the current time tc from the predicted time tp. The dialogue time estimation unit 50 estimates a time t2 required to generate a voice processing event of the low priority electronic mail service 404.

  The notification time estimation may be executed only when a predetermined condition is satisfied in order to reduce the calculation load. For example, the notification time estimation for the route guidance service may be executed only when a route is set in the car navigation system or during traveling.

  FIGS. 3A, 3B, and 3C show driving situations when the route guidance service 402 and the e-mail service 404 of the car navigation system are simultaneously used in the information processing apparatus or the electronic apparatus 10 of FIG. The forecast time of route guidance and route guidance is shown. The voice interaction system 12 performs voice notification by the route guidance service 402 with high priority and voice guidance and input instruction by the low priority electronic mail service 404 based on the prediction of the driving situation.

  In the driving situation on the upper side of FIG. 3A, the route guidance service 402 or the notification time estimation unit 20 obtains the route information (82) to the destination obtained from the car navigation system information storage unit 80, and the surrounding conditions on the map. According to (86), the current position (84), and the vehicle speed acquired from the vehicle speed sensor 112, as a driving situation with respect to the time axis, after traveling straight ahead for a long time, stop, turn right, and travel straight ahead Predict that. The notification time estimation unit 20 causes the route guidance service 402 to utter a high-priority voice notification “right” according to the service scenario immediately before stopping while driving, and a few seconds before that According to the service scenario, it is predicted that a high-priority voice notification “This is the right of XX is right” will be uttered. Dialog classification and priority are added to the voice notification service scenario of each voice notification of the route guidance service 402. In the estimated route guidance time on the lower side of FIG. 3 (A), the free time estimation unit 30 before the predicted time of the first voice notification of “This is the right of XX” indicated by the shaded area at the current time tc. An existing free time t1 is calculated.

  In the driving situation on the upper side of FIG. 3B, unlike the prediction, the traffic light at the pedestrian crossing turns red while the vehicle is running, and the driver stops the vehicle at the stop line and at the same time, the e-mail service 404. Start up. The route guidance service 402 or the notification time estimation unit 20 corrects the prediction of the driving situation according to the occurrence of the stop of the vehicle. Dialogue classification and priority are added to the email transmission service scenario of the email service 404. At this time, the vacant time estimation unit 30 calculates a vacant time t1 until the predicted time tp of the first voice notification “This is the right of XX” on the route guidance service 402 at the current time tc. The dialogue time estimation unit 50 uses the character number counter 54 and the input time estimation unit 52 to predict the number of voice guidance mora of “Where is the destination?” With a low priority of the e-mail service 404. The total predicted speech processing time t2 of the time, the key input or voice input (“...”) By the user as the answer and the estimated required time for performing the speech recognition is calculated. The input time estimation unit 52 may estimate the time by adding an additional time to the time corresponding to the maximum number of characters of voice input assumed from the dialogue history information of the corresponding scenario in the information storage unit 60.

  The dialogue control unit 70 compares the predicted time t2 of the low priority voice processing with the idle time t1 of the high priority voice processing. When the predicted time t2 is less than the free time t1 (t2 <t1), the dialogue control unit 70 utters a low-priority voice guidance “Where is the destination?” By the e-mail service 404, and then The user waits for a voice input via the acoustic device 106 or a key operation via the input device 108 by the user via the input processing unit 14 for a predetermined time secured in step (b). In the case of FIG. 3B, since the predicted time t2 is less than the idle time t1, the dialogue control unit 70 performs voice guidance, user voice input, and voice recognition using the input processing unit 12 and the output processing unit 14. Make it. If there is a voice input or key operation by the user, the input processing unit 72 performs input processing including voice recognition in accordance with it, and the dialog control unit 70 performs the next dialog control. On the other hand, when the estimated time t2 of the e-mail service is equal to or greater than the free time t1 (t2 ≧ t1), the route guidance service 402 first sends a voice notification with a high priority of “This is the right to the right”. Speak. Thereafter, the above-described processing is repeated until the predicted time t2 becomes less than the next free time t1.

  In the driving situation on the upper side of FIG. 3C, the driver resumes running after the above-mentioned stop. The route guidance service 402 or the notification time estimation unit 20 corrects the prediction of the driving situation according to the occurrence of the restart of the travel. While the vehicle is traveling, the driver operates the input device 108 in the e-mail service 404 so as to read out a low-priority received e-mail by voice. Dialogue classification and priority are added to the e-mail reading service scenario of the e-mail service 404. At this time, the vacant time estimation unit 30 predicts the vacant time t1 until the voice notification with the high priority “This is the right of XX is right”. The dialogue time estimation unit 50 uses the character counter 54 to calculate an estimated time t2 required for reading out the voice of the email based on the number of characters (number of mora) of the email.

  The dialogue control unit 70 compares the predicted time t2 of the low priority voice processing with the idle time t1 of the high priority voice processing. When the predicted time t2 of the electronic mail service is less than the free time t1 (t2 <t1), the dialogue control unit 70 causes the output processing unit 14 to read out the electronic mail. When the predicted reading time t2 of the e-mail is equal to or longer than the free time t1 (t2 ≧ t1), the dialogue control unit 70 suppresses the reading operation of the e-mail by the e-mail service 404 and is immediately or scheduled. At the predetermined time, the route guidance service 402 causes the next high-priority voice notification “This is the right of XX” to be uttered.

  At the time of the utterance, the new idle time t1 until the next “right direction” notification of the route guidance service 402 becomes the value of t1 ′ in FIG. The dialogue control unit 70 compares the estimated time t2 required for reading out the voice of the e-mail with the idle time t1. When the estimated time t2 of the electronic mail service is less than the free time t1 (t2 <t1), the dialogue control unit 70 causes the electronic mail service 404 to read out the electronic mail. When the predicted reading time t2 of the e-mail is equal to or longer than the free time t1 (t2 ≧ t1), the dialogue control unit 70 suppresses the reading operation of the e-mail by the e-mail service 404 and is immediately or scheduled. At a predetermined time, the route guidance service 402 causes the next high-priority voice notification “to the right” to be uttered. Thereafter, the same processing as described above is repeated until the predicted time t2 becomes less than the next free time t1.

  FIG. 4 shows a configuration of the information processing apparatus or electronic device 10 having the voice interaction system 12 that performs voice output and voice input of the fuel warning service 406 and the nearby facility search service 408 according to the embodiment of the present invention. In this case, the occurrence of a high priority warning by the fuel warning service 406 or the like is predicted based on the detection information of each sensor of the vehicle.

  The notification time estimation unit 20 includes a fuel reduction rate estimation unit 23 connected to the fuel sensor 114 and a fuel warning time estimation unit 24 connected to the fuel reduction rate estimation unit 23. The fuel warning service 406 periodically monitors the fuel sensor 114, and when the remaining amount of fuel falls below a predetermined threshold, a high priority voice “please refuel” according to the service scenario. Make a notification. Dialogue classification and priority are added to the voice warning service scenario of the fuel warning service 406. The fuel sensor 114 supplies the remaining amount of fuel to the fuel decrease rate estimation unit 23. The fuel reduction rate estimation unit 23 periodically stores the remaining amount of fuel and its time in a memory, and based on the change in the remaining amount of fuel at each periodic time, the remaining amount of fuel is a predetermined threshold value. When it becomes lower, the fuel decrease rate is calculated and supplied to the fuel warning time estimation unit 24. Based on the current remaining fuel amount and the fuel reduction rate, the fuel warning time estimation unit 24 warns the time when the fuel becomes less than another predetermined threshold and warns the necessity of fuel reduction or fuel replenishment. Is estimated. The neighborhood facility search service 408 searches the map database 86 or several nearby gas stations according to the user's request entered using the input device 108.

  Instead of or in addition to the fuel warning service 406, based on various sensors, oil change time warning, tire change time warning, front inter-vehicle distance warning, tire air pressure warning, emergency vehicle approach warning, pedestrian warning Voice services for motorcycle approach warnings, etc. may be provided.

  While the vehicle is traveling, the driver operates the input device 108 to search for a nearby facility with a low priority in the nearby facility search service 408. Dialog classification and priority are added to the service scenario for selecting a gas station in the neighborhood facility search service 408. The free time estimation unit 30 calculates the free time t1 until the predicted time tp of the voice notification “please replenish fuel” by the fuel warning service 406 at the current time tc. The neighborhood facility search service 408 displays a list of neighborhood gas stations as a search result and allows the user to select one. To that end, the neighborhood facility search service 408 predicts the estimated time required for the number of mora of low-priority voice guidance “Please select a gas station”, voice input by the user, and voice recognition thereof. The total predicted speech processing time t2 with the required time is calculated.

  The dialogue control unit 70 compares the predicted time t2 with the idle time t1. When the predicted time t2 is less than the free time t1 (t2 <t1), the dialogue control unit 70 utters a voice guidance “Please select a gas station” by the facility search service 408 in the neighborhood and is secured. The user waits for voice input or key input through the input processing unit 14 for a predetermined time. If there is an input by the user, the dialogue control unit 70 performs processing according to the input. When the estimated time t2 of the nearby facility search service 408 is equal to or greater than the free time t1 (t2 ≧ t1), the dialogue control unit 70 utters a voice guidance “Please select a gas station” and the user's Both of the inputs are suppressed, and a high-priority voice notification “please refuel” is uttered by the fuel warning service 406 immediately or at a timing to be warned. Thereafter, the above-described processing is repeated until the predicted time t2 becomes less than the next free time t1.

  The notification time estimation unit 20 further uses, for example, an inter-vehicle distance warning service based on the distance between the own vehicle detected by a millimeter wave radar or the like and another vehicle ahead or behind and the rate of decrease thereof. You may estimate the time to arrive at. The notification time estimation unit 20 further includes, for example, an oil change warning service and a tire friction warning service based on the travel distance from the date of the previous oil change and tire change to the current time and the vehicle speed, and the oil change warning and tire friction. The warning time may be estimated.

  The notification time estimation may be executed only when a predetermined condition is satisfied in order to reduce the calculation load. For example, the notification time estimation for the inter-vehicle distance warning service can be executed only when the inter-vehicle distance is within 100 m.

  FIG. 5 shows a configuration of an information processing apparatus or electronic device 10 having a voice dialogue system 12 that performs voice output and voice input of a driving time warning service 410 and a restaurant search service 412 according to still another embodiment of the present invention. Yes. In this case, based on the estimated or detected driver information, the occurrence of a high priority warning by the driving time warning service 410 or the like is predicted.

  The notification time estimation unit 20 includes a warning time estimation unit 26. The warning time estimation unit 26 is connected to an operation time timer 116 that measures the operation time. The operation time warning service 410 detects the operation time td that is continuously performed excluding the continuous engine stop time of, for example, less than 10 minutes, and when the threshold value thd is exceeded, “td time, “I am driving continuously. Please take a break.” Dialog classification and priority are added to the voice warning service scenario of the driving time warning service 410. The warning time estimation unit 26 compares the continuous operation time td received from the operation time timer 11 with the threshold thd seconds, and estimates the time tp at which continuous operation should be warned when it exceeds the threshold thd. The restaurant search service 412 searches several nearby restaurants on the map database 86 or the Internet in accordance with a user request input using the input device 108.

  Instead of or in addition to the driving time warning service 410, a voice service for a driver doze driving warning service, etc. may be provided. For example, for the dozing warning service, the notification time estimating unit 20 detects a physiological state with a driver biosensor, and estimates the time of falling asleep from a sign of dozing.

  While the vehicle is traveling, the driver operates the input device 108 to perform a restaurant search with a low priority in the restaurant search service 412. Dialogue classification and priority are added to the restaurant selection service scenario of the restaurant search service 412. The vacant time estimation unit 30 calculates the vacant time t1 up to the predicted time tp at which the voice notification “td time, continuous driving. Take a break.” By the driving time warning service 410 is given. The neighborhood facility search service 408 displays a list of nearby restaurants as a search result, and allows the user to select one. For this purpose, the restaurant search service 412 uses the character counter 54 based on the number of characters (number of mora) to determine the estimated time required for voice guidance “Please select a restaurant type” or “Please select a restaurant”. And a predicted speech processing time t2 that is the sum of the required time for speech input by the user and speech recognition determined by the input time estimation unit 52 based on the user response history information in the information storage unit 60 is calculated.

  The dialogue control unit 70 compares the predicted time t2 with the idle time t1. When the predicted time t2 is less than the vacant time t1 (t2 <t1), the dialogue control unit 70 utters a voice guidance “Please select the type of restaurant” by the restaurant search service 412 and the predetermined predetermined time is secured. Waiting for voice input or key input by the user via the input processing unit 14 for the time. If there is an input by the user, the dialogue control unit 70 performs processing according to the input. When the predicted time t2 of the restaurant search service 412 is equal to or greater than the vacant time t1 (t2 ≧ t1), the dialogue control unit 70 suppresses the voice processing of the restaurant search service 412 and immediately or at the timing to warn The warning service 410 first utters a high-priority voice notification “Td time, continuous operation. Take a break.” Thereafter, the above-described processing is repeated until the predicted time t2 becomes less than the next free time t1.

According to the above-described embodiment, for example, when the route guidance service 402 in FIG. 2 and the restaurant search service 412 in FIG. 5 are combined, the voice notification of the route guidance service 402 in FIG. When there is only a short free time before (t2 ≧ t1), when the voice query “What type of restaurant is desired?” Of the restaurant search service 412 in FIG. Notification and voice processing proceed as follows.
-(Deterred question: What type of restaurant would you like?)
-(Route guidance) This is the XX intersection.
-(Delayed Question) What kind of restaurant would you like?
-(User response) Chinese-(Confirmation) Chinese restaurant. Search
Accordingly, the voice processing of one voice processing unit (question-answer-confirmation notification) related to restaurant search is executed by the voice dialogue system 12 without interruption.

  FIG. 6 shows an information processing apparatus or electronic device 10 having a voice interaction system 12 that performs voice output and voice input of a route guidance service 402, an oil change warning service 414, and a news service 416 according to still another embodiment of the present invention. The structure of is shown. In this case, the route guidance service 402 predicts the highest priority voice notification based on the position of the vehicle and the surrounding environment obtained from the map database, web page or text traffic information. In addition, based on the detection information of each sensor of the vehicle, the occurrence of the next highest priority warning by the oil change warning service 414 or the like is predicted.

  The oil change warning service 414 periodically monitors the oil sensor 118 and, when the oil deterioration rate becomes higher than a predetermined threshold value, makes a voice notification “please change oil” according to the service scenario. . Dialogue classification and priority are added to the voice warning service scenario of the oil change warning service 414. The notification time estimation unit 20 includes a next guidance time estimation unit 22 shown in FIG. 1, an oil deterioration rate estimation unit 27 connected to the oil sensor 118, and an oil warning time estimation unit 28 connected to the oil deterioration rate estimation unit 27. Including. The oil sensor 118 supplies the oil deterioration rate to the oil deterioration rate estimation unit 27.

  The oil deterioration rate estimation unit 27 periodically stores the oil deterioration rate and its time in a memory, and the oil deterioration rate is a predetermined threshold value based on the change of the oil deterioration rate at each periodic time. When it becomes higher, the oil deterioration rate is supplied to the oil warning time estimation unit 28.

  While the vehicle is traveling, the driver operates the input device 108 in the news service 416 so as to read out the low priority news by voice. The news service 416 reads out the received text-format news in accordance with the input user request. Dialogue classification and priority are added to the voice service scenario of the news service 416. The oil warning time estimation unit 28 estimates a time tp1 at which warning of the necessity for oil replacement is to be made based on the current oil deterioration rate. For the route guidance service 402, as described with reference to FIG. 2, the next guidance time estimation unit 22 estimates the predicted time tp2 to be notified by voice based on the predicted driving situation. The available time estimation unit 30 calculates the current available time t1 = tp−tp by subtracting the current time tc from the earlier time tp = tp1 or tp2 of the predicted times tp1 and tp2. The dialogue time estimation unit 50 uses the character counter 54 to calculate a predicted time t2 required for reading the news by voice based on the number of news characters (number of mora).

  The dialogue control unit 70 compares the predicted reading time t2 of the news with the free time t1. When the predicted time t2 is less than the free time t1 (t2 <t1), the dialogue control unit 70 uses the output processing unit 14 to read the news.

  When the predicted time t2 is equal to or longer than the free time t1 (t2 ≧ t1), the dialogue control unit 70 suppresses reading of news by the news service 416 and immediately precedes the route guidance service 42 at a timing to be warned. The highest priority voice notification or the oil change warning service 414 next to the highest priority voice alert. Thereafter, the above-described processing is repeated until the predicted time t2 becomes less than the next free time t1.

  Further, there is an overlap between the required time period at the predicted time of the highest priority voice notification of the route guidance service 42 and the time period at the predicted time of the next highest priority voice warning after the oil change warning service 414. In this case, the dialogue control unit 70 compares the predicted time t2 of the second highest priority voice warning after the oil change warning service 414 with the highest priority voice notification idle time t1 of the route guidance service 42. When the predicted time t2 is less than the free time t1 (t2 <t1), the dialogue control unit 70 makes a warning by the oil change warning service 414. When the predicted time t2 is equal to or longer than the free time t1 (t2 ≧ t1), the dialogue control unit 70 suppresses the warning by the oil change warning service 414 and immediately or at the predetermined time before the route guidance service. The voice notification with the highest priority of 42 is performed.

  As described above, according to the embodiment of the present invention, it is possible to prevent contention between a plurality of voice input / output processes of each voice service with respect to a plurality of voice services being executed. Can be performed at an appropriate timing for the user without being interrupted at an inappropriate timing, and the user's thoughts and voice input or input operation are not hindered.

  Although the voice service has been described in the above-described embodiment, the present invention can also be applied to prevent contention between display output and operation input / output processing of different priorities for the user.

  The embodiments described above are merely given as typical examples, and it is obvious to those skilled in the art to combine the components of each embodiment, and variations and variations thereof will be apparent to those skilled in the art. Obviously, various modifications may be made to the above-described embodiments without departing from the scope of the invention as set forth in the scope.

Regarding the embodiment including the above examples, the following additional notes are further disclosed.
(Appendix 1) An information processing apparatus having a voice interaction function,
Service scenario storage means for storing a plurality of service scenarios having priorities;
Notification time estimating means for estimating the timing of the next first voice processing according to a service scenario with high priority in the service scenario storage means;
Free time estimation means for estimating free time from the current time to the timing of the first audio processing;
Dialogue time estimation means for estimating a time length required for the second voice processing including voice output in accordance with a low priority service scenario in the service scenario storage means;
Dialog control for performing the second voice processing in the idle time according to the low priority service scenario when the estimated duration of the second voice processing is less than the estimated idle time Means,
An information processing apparatus comprising:
(Supplementary note 2) When the estimated time length of the second voice processing is equal to or greater than the estimated idle time, the dialogue control means is configured to perform the first operation according to the low priority service scenario. 2. The information processing apparatus according to appendix 1, wherein the first voice processing is performed in accordance with the service scenario having a high priority while suppressing the second voice processing.
(Additional remark 3) Furthermore, the said dialog time estimation means includes the input time estimation means which estimates the time length required for a user's input, and the means which estimates the time to output voice based on the number of characters of a text. The information processing apparatus according to attachment 1, wherein the information processing apparatus is characterized.
(Supplementary Note 4) Further, an information storage unit is provided,
The dialogue time estimation means is a time required for the second voice processing by the low priority service scenario based on the dialogue history information, user response time history information or text data stored in the information storage unit. The information processing apparatus according to appendix 1, characterized in that
(Supplementary Note 5) The high priority service scenario is a route guidance service for a car navigation system,
The notification time estimation means estimates the timing of the first voice processing based on route information, current position information and a map database of the car navigation system. Information processing device.
(Supplementary Note 6) The supplementary note 1 is characterized in that the notification time estimation means is connected to a sensor and estimates the timing of the first voice processing based on detection information from the sensor. Information processing device.
(Supplementary note 7) The first voice processing includes voice notification, and the second voice processing includes voice question, voice recognition of a user's answer input voice, and voice confirmation of a reply. The information processing apparatus described in 1.
(Supplementary Note 8) A program for performing voice dialogue for an information processing apparatus having a service dialogue storage means for storing a plurality of service scenarios having classification and priority, and having a voice dialogue function,
Estimating the timing of the next first voice processing in accordance with a high priority service scenario in the service scenario storage means;
Estimating a free time from the current time to the timing of the first audio processing;
Estimating a length of time required for second audio processing including audio output according to a low priority service scenario in the service scenario storage means;
Performing the second audio processing in the idle time according to the low priority service scenario when the estimated duration of the second audio processing is less than the estimated idle time; ,
A program that can run to run.

FIG. 1 shows a configuration of an information processing apparatus or an electronic apparatus having a voice dialogue system that performs voice output and voice input of a plurality of voice services according to the present invention. FIG. 2 shows a configuration of an information processing apparatus or electronic device having a voice dialogue system for performing voice processing, that is, voice output and voice input of a route guidance service and an e-mail service of a car navigation system according to an embodiment of the present invention. ing. (A), (B), and (C) of FIG. 3 are predictions of driving situations when the route guidance service and the e-mail service of the car navigation system are simultaneously used in the information processing apparatus or electronic device of FIG. The estimated time of route guidance is shown. FIG. 4 shows a configuration of an information processing apparatus or an electronic apparatus having a voice dialogue system that performs voice output and voice input of a fuel warning service and a nearby facility search service according to an embodiment of the present invention. FIG. 5 shows a configuration of an information processing apparatus or electronic device having a voice interaction system for performing voice output and voice input of a driving time warning service and a restaurant search service according to still another embodiment of the present invention. FIG. 6 shows a configuration of an information processing apparatus or an electronic apparatus having a voice dialogue system that performs voice output and voice input of a route guidance service, an oil change warning service, and a news service according to still another embodiment of the present invention. Yes.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Information processing apparatus or electronic device 12 Voice dialog system 20 Notification time estimation part 30 Free time estimation part 40 Service scenario storage part 50 Dialog time estimation part 60 Information storage part 70 Dialog control part 72 Input processing part 74 Output processing part

Claims (6)

An information processing apparatus having a voice interaction function,
Service scenario storage means for storing a plurality of service scenarios having priorities;
Notification time estimating means for estimating the timing of the next first voice processing according to a service scenario with high priority in the service scenario storage means;
Free time estimation means for estimating free time from the current time to the timing of the first audio processing;
Conversation time estimation means for estimating a time length required for second voice processing including voice output processing and processing for accepting input by a user in accordance with a low priority service scenario in the service scenario storage means;
Dialog control for performing the second voice processing in the idle time according to the low priority service scenario when the estimated duration of the second voice processing is less than the estimated idle time Means,
An information processing apparatus comprising:   When the estimated time length of the second voice processing is equal to or greater than the estimated idle time, the dialogue control means is configured to enable the second voice processing according to the low priority service scenario. The information processing apparatus according to claim 1, wherein the first voice processing is performed according to the high priority service scenario. Further, the dialogue time estimation means includes an input time estimation means for estimating a time length required for input by a user, and a means for estimating a time for voice output based on the number of characters of text. The information processing apparatus according to claim 1 or 2 . It also has an information storage unit,
The dialogue time estimation means is a time required for the second voice processing by the low priority service scenario based on the dialogue history information, user response time history information or text data stored in the information storage unit. and characterized in that to estimate the information processing apparatus according to any one of claims 1 to 3. The high priority service scenario is a route guidance service for a car navigation system,
The notification time estimating means, the route information of the car navigation system, and characterized in that to estimate the timing of the first speech processing based on the current position information and the map database, according to claim 1 to 4 The information processing apparatus according to any one of the above. 6. The first voice processing includes voice notification, and the second voice processing includes voice question, voice recognition of a user's answer input voice, and voice confirmation of a reply. The information processing apparatus according to any one of the above.

Images