JP2017161813A - Response system and response program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique capable of reducing a possibility that a user's concentration is disturbed by an output of a response.SOLUTION: A response system includes: an input receiving part for receiving an input of a user; a response generating part for generating a response to the input; a steering state determining part for determining whether the user is steering or not; and a response outputting part for outputting the response when the user is not steering.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a response system and a response program.

  Conventionally, a voice recognition technique for recognizing a user's utterance such as a vehicle is known. For example, Patent Document 1 discloses a configuration in which voice input is performed when a vehicle state such as a radio wave condition does not satisfy a predetermined condition in a configuration in which voice is transmitted from a terminal to a voice recognition server to perform voice recognition. Is disclosed. Patent Document 2 discloses a configuration in which dictation is prohibited while a vehicle is running in a vehicle-mounted speech recognition device.

JP 2003-345391 A JP 2003-162295 A

In the above-described prior art, execution of speech recognition is prohibited under specific conditions (speech recognition is not started). However, even if the start timing of voice recognition is limited, there are few merits for the user. For example, utterances that serve as triggers for starting speech recognition under various conditions during driving rarely hinder user concentration. On the other hand, when a user who is taking action that needs to respond to changes in the situation, such as driving a vehicle, starts outputting a voice recognition result or a response based on the voice recognition result, the user drives In addition to this, it is necessary to concentrate on the recognition of responses, and the burden on the user may become excessive.
The present invention has been made in view of the above problems, and an object of the present invention is to provide a technique capable of reducing the possibility that the concentration of users is hindered by the output of responses.

  In order to achieve the above object, an input receiving unit that receives user input, a response generation unit that generates a response to the input, a steering state determination unit that determines whether or not the user is steering, and use When the person is not steering, a response system including a response output unit that outputs a response is configured.

  Furthermore, in order to achieve the above object, the computer includes an input receiving unit that receives a user's input, a response generating unit that generates a response to the input, and a steering state determination unit that determines whether or not the user is steering A response program is configured to function as a response output unit that outputs a response when the user is not steering.

  That is, the response system and the response program output a response to the user's input when the user is not steering. When steering, the user should concentrate on driving, and when a response to an input (for example, a display that requires gaze on the screen or a sound output that should be listened carefully) is executed, the user responds May not be properly recognized. Therefore, if the response is output when the vehicle is not being steered, it is possible to reduce the possibility of the user's concentration being hindered by the output of the response.

It is a block diagram which shows a navigation system. FIG. 2A is a flowchart showing response processing, and FIG. 2B is a flowchart showing external response generation processing. FIG. 3A is a diagram showing an example of postponing response output, and FIG. 3B is a diagram showing an example of output after postponement.

Here, embodiments of the present invention will be described in the following order.
(1) Configuration of navigation system:
(2) Response processing:
(3) External response generation processing:
(4) Other embodiments:

(1) Configuration of navigation system:
FIG. 1 is a block diagram showing a configuration of a navigation system 10 according to an embodiment of the present invention. The navigation system 10 includes a control unit 20 including a CPU, a RAM, a ROM, and the like, and the control unit 20 can execute a desired program recorded in the ROM or the local storage medium 30. In this embodiment, a navigation program can be executed as one of the programs. The navigation program can cause the control unit 20 to realize a function of displaying a current location and a map around the current location on the display unit and a function of searching and guiding a route to the destination.

  In addition, the navigation program includes a response program, and the control unit 20 executes a function of performing a voice conversation between the user and the navigation system 10 by the processing of the response program 21. The dialogue is realized by executing the input by the user's uttered voice and the response from the navigation system 10 at least once.

  A vehicle in which the navigation system 10 is used includes a GPS receiver 41, a vehicle speed sensor 42, a gyro sensor 43, a user I / F unit 44, a communication unit 45, and a steering sensor 46. The GPS receiver 41 receives radio waves from GPS satellites and outputs a signal for calculating the current location of the vehicle via an interface (not shown). The vehicle speed sensor 42 outputs a signal corresponding to the rotational speed of the wheels provided in the vehicle. The control unit 20 acquires this signal via an interface (not shown) and acquires the vehicle speed. The gyro sensor 43 detects angular acceleration about turning in the horizontal plane of the vehicle, and outputs a signal corresponding to the direction of the vehicle.

  The control unit 20 acquires this signal and acquires the traveling direction of the vehicle. The control unit 20 acquires the current location of the vehicle by specifying the travel locus of the vehicle based on output signals from the vehicle speed sensor 42 and the gyro sensor 43 and the like. The output signal of the GPS receiver 41 is used for correcting the current location of the vehicle specified by the vehicle speed sensor 42, the gyro sensor 43, and the like.

  The user I / F unit 44 is an interface unit for providing various information to the user and receiving various inputs from the user, and includes a display unit, an operation input unit, a speaker, a microphone, and the like (not shown). ing. The control unit 20 can display, on the display unit, a map and route around the current location of the vehicle, a facility search result, and the like by the function of the navigation program. Moreover, the control part 20 can acquire a user's speech sound with a microphone, and can output arbitrary sounds including a sound from a speaker based on various sound data.

  The communication unit 45 includes a circuit for wirelessly communicating with the external system 50, and arbitrary information is exchanged between the control unit 20 and the external system 50 via the communication unit 45. In the present embodiment, the control unit 20 can transmit information indicating the user's uttered voice to the external system 50 via the communication unit 45.

  The external system 50 includes an external storage medium (not shown), and an external database 50a is recorded on the external storage medium. When the external system 50 acquires information indicating the uttered voice, the external system 50 performs voice recognition based on the external database 50a, and generates an external response that is a response to the character string as a recognition result. When the external response is generated, the external system 50 transmits information indicating the external response to the navigation system 10. The control unit 20 can receive information indicating the external response via the communication unit 45.

  In the present embodiment, the external database 50a includes information that can be updated and information that can be increased over time. For example, the external system 50 performs a process of collecting various news information, weather information of each place, traffic information of an arbitrary road, etc., and records and manages the collected information as an external database 50a. The external system 50 can search for information to be searched from the external database 50a and generate it as an external response.

  The steering sensor 46 is a sensor that outputs the rotation angle of the steering included in the vehicle, and the control unit 20 can acquire the steering angle based on the output signal of the steering sensor 46. The steering angle may be acquired as, for example, an angle in which the steering operation is not performed at 0 °, a specific rotation direction is positive, and the reverse rotation direction is negative.

  A local database 30a is recorded in the local storage medium 30. The local database 30a includes map information 30a1 and a voice database 30a2. The map information 30a1 is information used for specifying the current location, searching for facilities, and the like, node data indicating the positions of nodes set on the road, and shape interpolation points for specifying the shape of the road between the nodes. It includes data, link data indicating the connection between nodes, facility data indicating the location, attributes, names, etc. of facilities existing on and around the road.

  Therefore, the facility data is data in which records for each of a plurality of facilities are configured, and information for each of a plurality of items such as a position, an attribute, and a name is associated with each record. In the present embodiment, the attributes of each facility are associated with information indicating the type of facility, for example, a restaurant, a convenience store, a public facility, and the like.

  The voice database 30a2 is information that is referred to when a user's uttered voice is recognized and converted into a character string. The voice database 30a2 can be various information according to an algorithm for executing voice recognition. For example, the control unit 20 creates a phoneme string by examining a feature amount corresponding to a phoneme from the waveform of a speech voice, and obtains a character string that is likely to correspond to the phoneme string by comparing the phoneme string with an acoustic model and a language model. When having the configuration, the voice database 30a2 includes information indicating the acoustic model and the language model.

  The response program 21 includes an input reception unit 21a, a response generation unit 21b, a steering state determination unit 21c, and a response output unit 21d in order to output a response to a user input. The input receiving unit 21a is a program module that causes the control unit 20 to realize a function of receiving user input. In the present embodiment, a configuration is adopted in which the control unit 20 specifies the content that the user wants to respond to based on the user's utterance voice, and the control unit 20 uses the microphone to utter the user's utterance voice. To get.

  The response generation unit 21b is a program module that causes the control unit 20 to realize a function of generating a response that is a response to an input. That is, the control unit 20 performs voice recognition on the uttered voice received by the process of the input receiving unit 21a by the process of the response generation unit 21b, and generates a response based on the recognition result of the uttered voice. Generate a local response. In addition, the control unit 20 causes the external system 50 to perform speech recognition of the uttered speech, and generates an external response based on the recognition result. The local response and the external response are combined and regarded as a response system response.

  Various modes of input and response can be assumed. For example, an example in which the input and response are a request for information and presentation of information, or an example in which an instruction and execution are performed is assumed. In the present embodiment, an example is assumed in which the input is a search request (designation of search conditions) and the response is an output of a search result. In the speech recognition, for example, the control unit 20 creates a phoneme string by examining the feature amount corresponding to the phoneme from the waveform of the uttered speech accepted by the processing of the input accepting unit 21a, and the acoustic model recorded as the speech database 30a2. This can be realized by comparing the language model with the phoneme sequence to obtain a character string that is likely to correspond to the phoneme sequence. When the character string is acquired, the control unit 20 specifies a search target and a search condition from the character string, refers to the local database 30a, and searches for the search target based on the search condition.

  On the other hand, the control unit 20 transmits information indicating the speech voice to the external system 50 via the communication unit 45. When the external system 50 acquires information indicating the uttered voice, the external system 50 refers to the external database 50a and recognizes the uttered voice. The voice recognition can be realized by the same processing as the recognition in the control unit 20. When voice recognition is performed and a character string is specified, the external system 50 specifies a search target and a search condition indicated by the character string, and acquires a search result by referring to the external database 50a. When the search result is obtained, the external system 50 transmits the search result to the navigation system 10, and the control unit 20 acquires the search result via the communication unit 45.

  As a result of the above processing, a response is generated by the local database 30a and the external database 50a. Therefore, the control unit 20 combines the two by deleting duplicates of the search results, and the response to the input. To do. Since the external system 50 can communicate with a plurality of navigation systems 10, if the maintenance of the external database 50 a is performed, the effect can be enjoyed by the plurality of navigation systems 10.

  For this reason, the external database 50a includes data for recognizing words indicating search targets and search items that are not included in the local database 30a from the character string with high probability. Therefore, by using voice recognition and response generation by the external system 50 together, there is a high possibility that more advanced recognition and response generation will be performed. As described above, in a configuration in which a response is generated using not only the local database 30a but also the external database 50a, a time lag that can be recognized by the user from when the user inputs to when the response is generated. In some cases, the case where the user is steering is likely to occur when the generation of the response is completed. Therefore, in this embodiment in which a response is generated using the external database 50a, an advantage of adopting a configuration that does not output a response during steering tends to appear.

  However, it is not essential that both the local database 30a and the external database 50a generate voice recognition and a response based on the recognition result, and voice recognition and response generation are performed by either or both of the databases. Good. For example, a configuration in which the control unit 20 determines a database to be referred to according to various rules may be employed. Of course, if only one database is referred to, the other database may be omitted.

  Here, a configuration is assumed in which speech recognition is performed based on both the local database 30a and the external database 50a, and a response to the recognition result is generated based on both the local database 30a and the external database 50a. . Of course, it is possible that information for creating a response to the recognition result exists only in one of the local database 30a and the external database 50a. In this case, a response is generated based on only one of the databases.

  For example, when the user's input is a facility search or the like, the facility can exist in both the local database 30a (map information 30a1) and the external database 50a, so that a response can be generated based on both. On the other hand, when the user input is a request for news information or the like, news information that can change over time exists in the external database 50a and does not exist in the local database 30a, so a response is made based only on the external database 50a. Can be generated.

  For example, when the character string is “tell me today's news”, the control unit 20 refers to the local database 30a to search for today's news. In the present embodiment, since today's news is not recorded in the local database 30a, a search result cannot be obtained from the local database 30a.

  In the external system 50, today's news is searched with reference to the external database 50a. In this embodiment, since today's news is managed in the external database 50a, a search result is obtained and transmitted to the navigation system 10. Of course, when it is clear that the search target is not recorded in the local database 30a or the external database 50a, the search processing in at least one of the databases may not be executed.

  The steering state determination unit 21c is a program module that causes the control unit 20 to realize a function of determining whether or not the user is steering. That is, the control unit 20 acquires the output signal of the steering sensor 46 and acquires the steering angle of the steering. Then, when the absolute value of the steering angle is not less than or equal to the threshold value, it is considered that the vehicle is being steered.

  The response output unit 21d is a program module that causes the control unit 20 to realize a function of outputting a response when the user is not steering. That is, if a response is output when the user is steering, there is a high possibility that the user's concentration will be hindered, so the control unit 20 does not output a response during steering and is not in steering. Output response at. Specifically, referring to the determination result obtained by the process of the steering state determination unit 21c, when the steering is in progress, the control unit 20 stands by without outputting a response. When the determination result by the process of the steering state determination unit 21c does not indicate that the steering is being performed, the control unit 20 causes the user I / F unit 44 to output the response generated by the process of the response generation unit 21b.

  Various modes of response can be assumed, and in the present embodiment, the control unit 20 outputs the voice that is the utterance sound of the character string to the user I / F unit 44 based on the character string indicating the response. Let Therefore, the control unit 20 outputs information indicating the voice that is the utterance sound of the response to the speaker of the user I / F unit 44. As a result, sound indicating a response is output from the speaker of the user I / F unit 44.

  According to the above configuration, no response is output during steering. Therefore, it is possible to reduce the possibility that the user's concentration is hindered by the output of the response, and the user can concentrate on driving. Also, since the response is output when the user is not steering, the user can listen carefully to the response even during driving, and can easily recognize and analyze the contents. .

(2) Response processing:
Next, response processing executed by the navigation system 10 will be described. In the navigation system 10, the response process shown in FIG. 2A is started by various triggers such as a user operation and detection of a predetermined word by a microphone. In the response process, the control unit 20 receives an input by the uttered voice by the process of the input receiving unit 21a (step S100). That is, the control unit 20 acquires waveform information indicating the user's uttered voice via the microphone of the user I / F unit 44. FIG. 3A shows an example in which the user U utters “Tell me today's news”.

  Next, the control unit 20 transmits the uttered voice to the external system 50 by the process of the response generation unit 21b (step S105). That is, the control unit 20 transmits the waveform information indicating the uttered voice acquired by the process of step S <b> 100 to the external system 50 via the communication unit 45.

  Next, the control unit 20 performs voice recognition based on the local database 30a by the processing of the response generation unit 21b (step S110). That is, the control unit 20 refers to the voice database 30a2, analyzes the uttered voice acquired by the process of step S100 using a predetermined voice recognition algorithm, and acquires a character string having a high probability as the content of the uttered voice. Here, it is assumed that “tell me today's news” is recognized as a character string in the example shown in FIG. 3A.

  Next, the control part 20 acquires a search result based on the local database 30a by the process of the response generation part 21b (step S115). That is, the control unit 20 specifies a search target and a search condition from the character string acquired in the process of step S110. For example, in the example shown in FIG. 3A, since the character string “Tell me today's news” is obtained as a speech recognition result, the control unit 20 uses the news as a search target and associates today's date with it. Is a search condition.

  Then, the control unit 20 refers to the local database 30a and searches for a search target that matches the search condition. In the present embodiment, no news information is recorded in the local database 30a, so no news is searched. If it is previously specified that news information is not recorded in the local database 30a, the search may be omitted in step S115.

  Next, the control unit 20 determines whether or not an external response has been received by the processing of the response generation unit 21b (step S120). In other words, when the uttered voice transmitted in step S105 is received by the external system 50, an external response generation process is performed in the external system 50. When the external response is generated, information indicating the external response is transmitted from the external system 50 to the navigation system 10. Therefore, the control unit 20 determines whether or not the external response transmitted by the external system 50 has been acquired via the communication unit 45. If it is not determined in step S120 that an external response has been received, the control unit 20 waits until it is determined that an external response has been received. In the example shown in FIG. 3A, today's news is searched by the external system 50, and news of a predetermined number of items is extracted and used as an external response.

  Next, the control part 20 produces | generates a response by the process of the response production | generation part 21b (step S125). That is, the search result acquired in step S115 and the search result acquired from the external system 50 by the process of step S120 are combined. At this time, the control unit 20 deletes duplicate results in both search results. In the example shown in FIG. 3A, since the search result is not acquired in step S115, the control unit 20 regards the search result acquired from the external system 50 by the process of step S120 as a response.

  Next, the control unit 20 determines whether or not steering is being performed by the processing of the steering state determination unit 21c (step S130). That is, the control unit 20 refers to the output signal of the steering sensor 46 and acquires the current steering rotation angle. And if the absolute value of the said rotation angle is larger than the threshold value which can be considered that the steering is not rotating, the control part 20 will determine with it being steering. If it is not determined in step S130 that the vehicle is steering, the control unit 20 repeats the determination in step S130. That is, when the user is steering in the vehicle, the control unit 20 postpones the output of the response until the steering is completed. 3A shows a screen example 44a of the display unit of the user I / F unit 44. A road is indicated by a line, a planned route is indicated by a thick line, a vehicle is indicated by C, and a destination is indicated by G. In this example, the user of the vehicle C is steering to turn right. Therefore, when the vehicle C exists at the position shown in FIG. 3A, the output of the response is postponed.

  In the period in which the output of the response is postponed until the steering is finished, the control unit 20 may present to the user that the output of the response is postponed. In this case, the control unit 20 specifies information indicating that the output of the response has been postponed by selecting a predetermined character string or image, and outputs a control signal for outputting the information to the user I / F unit. 44 for output. As a result, the user I / F unit 44 outputs information indicating that the output of the response has been postponed. In FIG. 3A, the information indicating that the output of the response has been postponed is the character string “read after right turn”, and the utterance sound of the character string is output from the speaker of the user I / F unit 44 Is shown.

  On the other hand, when it determines with it being steering in step S130, the control part 20 outputs a response by the process of the response output part 21d (step S135). That is, the control unit 20 outputs a control signal for outputting the character string indicating the response generated in step S125 to the user I / F unit 44. As a result, the user I / F unit 44 outputs a character string indicating a response by voice. FIG. 3A shows an example in which “read today's news”, which is the first sentence composed of a character string indicating a response, is output. After this sentence, a voice indicating the content of the news is output.

(3) External response generation processing:
Next, an external response generation process executed by the external system 50 will be described. The external system 50 executes the external response generation process shown in FIG. 2B when a transmission request for the spoken voice is made from the navigation system 10. When the external response generation process is started, the external system 50 receives the input content (step S200). That is, the external system 50 acquires information indicating the speech voice transmitted from the navigation system 10 by a communication unit (not shown).

  Next, the external system 50 performs voice recognition based on the external database 50a (step S205). That is, the external system 50 refers to the external database 50a, analyzes the utterance voice indicated by the information received in step S200 with a predetermined voice recognition algorithm, and obtains a character string having a high probability as the content of the utterance voice. To do. In the example shown in FIG. 3A, when the uttered voice is transmitted to the external system 50, “tell me today's news” is recognized as a character string.

  Next, the external system 50 acquires a search result based on the external database 50a (step S210). That is, the external system 50 refers to the external database 50a, and specifies a search target and a search condition from the character string acquired in the process of step S205. For example, in the example shown in FIG. 3A, since a character string “tell me today's news” is obtained as a speech recognition result, the control unit 20 uses news as a search target and associates today's date with it. This is the search condition.

  Then, the control unit 20 refers to the external database 50a and searches for a search target that matches the search condition. In the present embodiment, since news information is recorded in the external database 50a, news that matches the search condition is searched. Therefore, the external system 50 acquires a predetermined number of items of news from the obtained search results and uses it as an external response. When the external response is generated, the external system 50 transmits the external response to the navigation system 10 (step S215). That is, the external system 50 transmits information indicating the external response acquired by the process of step S210 to the navigation system 10 via the communication unit.

(4) Other embodiments:
The above embodiment is an example for carrying out the present invention, and various other embodiments can be adopted as long as a response is output when the user is not steering. For example, the moving body that moves together with the navigation system 10 is arbitrary, and may be a vehicle or a pedestrian, and various examples are assumed. The response system may be a device mounted on a vehicle or the like, a device realized by a portable terminal, or realized by a plurality of devices (for example, a client and a server). It may be a system.

  Furthermore, at least a part of the input reception unit 21a, the response generation unit 21b, the steering state determination unit 21c, and the response output unit 21d constituting the response system may be divided into a plurality of devices. For example, after the controller 20 performs voice recognition based on the voice database 30a2, a character string after voice recognition is transmitted to the external system 50, and an external response generated in the external system 50 is acquired, or installed in a vehicle For example, the ECU may be configured to determine the steering state. Of course, a part of the configuration of the above-described embodiment may be omitted, and the processing order may be changed or omitted. For example, step S105 may be executed in parallel with the processing after step S110, or may be executed after step S120.

  The input reception part should just be able to receive a user's input. That is, the response system is a system that executes a response to the user's input, and the input receiving unit only needs to be able to receive the user's input so that the user can specify the content that the user wants to respond to. Various modes can be assumed for the input. For example, the input may be realized by the user operating the operation unit, or may be realized by acquiring the user's uttered voice using a microphone or the like.

  The response generation unit only needs to be able to generate a response to the input. That is, since the user's input content is content that requests a response to the response system, the response generation unit identifies the input content and executes a process for generating a response according to the input content. The specification of the input content can be realized by various methods according to the input mode. For example, when the input is selection of an option by operating the operation unit, input of a command, or the like, the input content can be specified by specifying the selected option and specifying the input command. When the user's uttered voice is an input, the input content can be specified by performing voice recognition and converting the voice into a character string.

  Generation of a response according to the input content can be realized by executing various processes according to the input content. For example, if the input content is specified by options and commands, it is only necessary to predetermine processing to be executed for each option and command. Of course, in each process, the process may be executed using parameters input by the user. Further, when the user's uttered voice is an input, a process to be executed may be specified by analyzing the contents converted into a character string, and the process may be executed. For example, in the case of the character string “the traffic information ahead”, the traffic information about the road ahead and the planned route is acquired.

  Various modes of input and response can be assumed. For example, an example in which the input and response are a request for information and presentation of information, or an example in which an instruction and execution are performed is assumed. More specific examples of the case where the input and response are information request and information presentation include an example of information request that matches the search condition and information showing the search result. Of course, various types of information can be assumed for the information to be provided, such as facility information, weather information, traffic information, information on newly opened stores, store price ranges, word-of-mouth reviews, and recommended stores. Various information may be assumed. As a more specific example in the case where the input and response are an instruction and an execution, a command for reproducing a file stored in a storage medium (reproduction of music, moving images, etc.) and its reproduction can be mentioned.

  The steering state determination unit only needs to be able to determine whether or not the user is steering. That is, the steering state determination unit determines whether the user can output the output of a sensor (for example, a steering angle sensor, an acceleration sensor, or a position sensor) that detects the steering state or a change in the vehicle that can estimate the change in the steering state. What is necessary is just to be able to determine whether it is steering.

  The period during which the steering should be performed may be a period during which the user should not concentrate on the response. In addition to the period during which the user is actually operating the steering, the period during which the user is preparing for steering, It may be a period estimated to be released. As a period during which preparation for steering is being performed or a period during which steering is estimated to be performed, for example, a lane change or traveling in a non-straight direction is performed depending on a blinking period, a planned travel route, or the like. Estimated period, parking operation, etc., period of lane change, driving in a non-straight direction, parking operation, etc., depending on the user's line of sight and face, etc. .

  The response output unit only needs to be able to output a response when the user is not steering. That is, if a response is output when the user is steering, there is a high possibility that the user's concentration is hindered, so the response output unit does not output a response during steering and is not in steering Output response at. Various modes of response output can be assumed, and information output, command execution, and the like may be used. Examples of the former include an example in which a display on an output unit such as a display or an audio output from an output unit such as a speaker is a response output. As the latter, for example, music is output by an output unit such as a speaker. Can be mentioned.

  Furthermore, the response generation unit may be configured to perform re-recognition by changing a parameter for recognizing the speech voice from the parameter in the initial recognition when the user is steering. In general speech recognition, there can be a plurality of estimated character string candidates that are uttered speech, and a character string that is most likely as a speech speech character string is obtained as a recognition result. Therefore, if the parameters for performing speech recognition are changed, the most likely character string can change.

  In general, the recognition rate is changed by adjusting the parameters. However, if high-accuracy recognition is enabled, it takes time to complete the recognition. Therefore, when the response output is postponed, if the configuration is such that the speech recognition speech is re-recognized by changing the initial recognition parameters so that more accurate recognition can be performed, Can output a response at an early stage in a state in which high speed is more important than accuracy. On the other hand, in a situation where steering is in progress, it is possible to acquire information in which accuracy is more important than high speed and output the information after steering.

  The parameters may be variable values, conditions, information, and the like that can be adjusted so as to improve the recognition rate. For example, the data amount and weight of the acoustic model or language model to be referred to may be adjusted. The variable parameters (beam width, insertion penalty, etc.) may be adjusted by the speech recognition algorithm. In the embodiment described above, the amount of data referred to in the audio database 30a2 or the external database 50a may be variable, or the control unit 20 or the external system 50 may change variable parameters such as beam width and insertion penalty. You may adjust.

  Note that the parameter change may be performed within the same device, or may be performed by changing the device. As the latter, for example, in the above-described embodiment, the control unit 20 initially performs voice recognition based on the local database 30a, does not perform voice recognition in the external system 50, and also uses the external database 50a during steering. Thus, the voice recognition is performed. The parameters may be set so that the voice recognition based on the external database 50a in the external system 50 is more accurate than the voice recognition in the local database 30a in the control unit 20.

  Further, when the user is steering, the response generation unit may regenerate the response, and the response output unit may output the regenerated response. That is, when the output of the response is postponed due to the steering, the value of the existing response may be reduced while the postponement is made, or the response content may be inaccurate or inappropriate. Therefore, if the response is regenerated when the vehicle is being steered, it is possible to prevent a decrease in the value of the response, increase accuracy and appropriateness, and the like.

  This configuration includes, for example, a configuration in which the control unit 20 returns to step S105 or step S115 and continues processing when it is determined in step S130 that steering is in progress. Note that the regeneration of the response may be executed under the same conditions as the initial response, or the response may be regenerated by setting a condition different from the initial condition.

  As the latter, for example, in a configuration in which the response indicates a search target search result indicated by an input, a configuration in which the regenerated response is a search result of a re-search in which the original search condition is changed may be mentioned. The search condition only needs to be changed according to a change in the situation during steering. For example, when traffic information on the road ahead is a response, the vehicle turns right or left due to steering, and the road ahead changes. In addition, there is a configuration in which traffic information on the road ahead after the change is retrieved and used as a response.

  This configuration can be realized, for example, by modifying the above-described embodiment. Specifically, when it is determined in step S130 that the vehicle is being steered, the control unit 20 identifies a change in the traveling direction of the vehicle based on the steering angle, and travels after the change in the traveling direction based on the map information 30a1. Estimate the road that will be. And the control part 20 communicates with the external system 50 via the communication part 45, acquires the traffic information about the estimated road, and regards it as a response.

  Of course, when the response is regenerated, various changes can be made in addition to the change in the condition corresponding to the situation change during steering. For example, in the search, a configuration for changing the search range can be cited. When changing the search range, for example, the search range of the search target may be changed (for example, the traffic information is searched for a wider area than the initial area), or the search target may be changed ( For example, a music file is searched by an initial search, and related information (artist related information, etc.) of a music file is searched by a re-search. Various structures can be employed by searching a database and searching an external database by re-search.

  Further, when the user is steering, the response generation unit acquires information related to input, and when the user is steering, the response output unit outputs information related to input. Good. That is, when the user is steering, information related to the input may be output in addition to the response corresponding to the input. The information related to the input includes, for example, information other than the information directly requested by the user input. More specifically, when a user inputs an instruction to play a music file, the music file is played, and information related to the music file (for example, videos and news about the artist) is searched. The configuration that is output as a response.

  In this configuration, for example, steps S105, S115, and S120 are omitted in the above-described configuration, and the control unit 20 specifies a music file to be played back in step S125 based on the result recognized in step S110. It is assumed that the music file to be played is played back in step S135. In this configuration, when it is determined in step S130 that the vehicle is being steered, the control unit 20 transmits the uttered voice to the external system 50, and the external system 50 relates to the music file based on the external database 50a. This can be realized by a configuration in which the information is searched for an external response, the control unit 20 acquires the external response, and outputs the external response in step S135.

  Further, the method of outputting a response when the user is not steering as in the present invention can be applied as a program or a method. The system, program, and method as described above can be realized as a single device or can be realized as a plurality of devices, and include various aspects. For example, it is possible to provide a navigation system, method, and program provided with the above means. Further, some changes may be made as appropriate, such as a part of software and a part of hardware. Furthermore, the invention can be realized as a recording medium for a program for controlling the system. Of course, the software recording medium may be a magnetic recording medium, a magneto-optical recording medium, or any recording medium to be developed in the future.

  DESCRIPTION OF SYMBOLS 10 ... Navigation system, 20 ... Control part, 21 ... Response program, 21a ... Input reception part, 21b ... Response generation part, 21c ... Steering state determination part, 21d ... Response output part, 30 ... Local storage medium, 30a ... Local database , 30a1 ... map information, 30a2 ... voice database, 41 ... receiving unit, 42 ... vehicle speed sensor, 43 ... gyro sensor, 44 ... user I / F unit, 45 ... communication unit, 46 ... steering sensor, 50 ... external system, 50a ... external database

Claims (7)

An input reception unit for receiving user input;
A response generator for generating a response to the input;
A steering state determination unit that determines whether or not the user is steering;
If the user is not steering, a response output unit for outputting the response;
A response system comprising: The input is
The spoken voice of the user,
The response is
Generated based on the recognition result of the speech,
The response system according to claim 1. The response generator is
When the user is steering, the parameter for recognizing the uttered voice is changed from the parameter in initial recognition, and re-recognition is performed.
The response system according to claim 2. The response generator is
If the user is steering, regenerate the response,
The response output unit
Output the regenerated response,
The response system according to any one of claims 1 to 3. The response is
Indicates the search result of the search target indicated by the input,
The regenerated response is
The search result of the re-search with the original search condition changed.
The response system according to claim 4. The response generation unit obtains information related to the input when the user is steering,
The response output unit
If the user is steering, output information related to the input;
The response system according to any one of claims 1 to 5. Computer
An input reception unit for receiving user input,
A response generator for generating a response to the input;
A steering state determination unit that determines whether or not the user is steering;
A response output unit that outputs the response when the user is not steering;
As a response program to function as.

Images