INFORMATION MANAGEMENT SYSTEM AND INFORMATION MANAGEMENT METHOD |
|||||||
申请号 | EP16855519.1 | 申请日 | 2016-10-14 | 公开(公告)号 | EP3364409A1 | 公开(公告)日 | 2018-08-22 |
申请人 | Yamaha Corporation; | 发明人 | IWATA, Takahiro; SETO, Yuki; OCHI, Yumiko; ISHIDA, Tetsuro; MORIGUCHI, Shota; IWASE, Hiroyuki; | ||||
摘要 | An information management system is a system for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, the system including: a text identifier configured to identify from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and an information generator configured to generate distribution information that indicates to the terminal device the related information corresponding to a modified text that is partially different from the registered text identified by the text identifier. | ||||||
权利要求 | |||||||
说明书全文 | The present invention relates to a technique for managing information provided for users. Conventionally, there has been proposed a technique for providing users with a variety of information in accordance with results of speech recognition of a spoken voice. In Patent Document 1, for example, there is disclosed a configuration in which speech recognition is carried out on a spoken voice of a user, and an area or a scale or the like is specified of a map displayed around a target position identified as a result of the speech recognition. Patent Document 1: It would be convenient for a hearing-impaired person who has difficulty hearing a guidance voice, or for a foreigner who does not understand the language of the guidance voice, if speech recognition could be performed on a guidance voice broadcast on a public transportation service such as a train or a bus, so that guidance information in the form of text, or in the form of translated text could be provided to a terminal device of a passenger. In reality, however, it is difficult to realize flawless speech recognition. Errors in recognition may occur as a result of, for example, pronunciation traits (habits) unique to individual speakers, or as a result of background noise or the like during voice receipt. When a recognition error occurs, there is a possibility that appropriate related information corresponding to the guidance voice may not be provided to users. In the foregoing, while reference is made to speech recognition of a guidance voice, similar problems may occur, for example, in a case where a person acting as a guide inputs a text represented by guidance voice. For example, in a case where a text entered by a guide person or a translation text thereof is presented as related information, there is a possibility that appropriate related information corresponding to the guidance voice may not be provided to users due to input error made by the guide person or for other reasons. In view of the foregoing, it is an object of the present invention to provide to users appropriate related information corresponding to a guidance voice. To solve the above-mentioned problems, an information management system according to a preferred aspect of the present invention is an information management system for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, and includes: a text identifier configured to identify from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and an information generator configured to generate distribution information that indicates to the terminal device the related information corresponding to a modified text that is partially different from the registered text identified by the text identifier. An information management system according to another aspect of the present invention is an information management system for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, and includes: a text identifier configured to identify from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and an information generator configured to generate distribution information that indicates to the terminal device the registered text identified by the text identifier, each of the registered texts including an insertion section for insertion of a selected one of multiple insertion phrases, with the text identifier being configured to execute a first process of identifying from among the multiple registered texts a registered text that is similar to the input text, and a second process of searching among the multiple insertion phrases for an insertion phrase corresponding to the input text, and the information generator being configured to, in a case where an insertion phrase corresponding to the input text is identified at the second process, generate distribution information that indicates the related information corresponding to a text obtained by inserting the insertion phrase identified at the second process into the insertion section of the registered text identified at the first process, whereas to generate distribution information that indicates the related information corresponding to a text corresponding to the registered text from which the insertion section has been removed in a case where an insertion phrase corresponding to the input text is not identified at the second process. An information management system according to another aspect of the present invention is an information management system for generating related information related to a guidance voice, and includes: a text identifier configured to identify from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and an information generator configured to generate the related information corresponding to a modified text that is partially different from the registered text identified by the text identifier. To solve the above-mentioned problems, an information management method according to a preferred aspect of the present invention is an information management method for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, and includes: identifying from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and generating distribution information that indicates to the terminal device the related information corresponding to a modified text that is partially different from the registered text identified by the text identifier. An information management method according to another aspect of the present invention is an information management method for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, and includes: identifying from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and generating the distribution information that indicates to the terminal device the registered text identified by the text identifier, each of the registered texts including an insertion section for insertion of a selected one of multiple insertion phrases. The identifying of a registered text includes executing a first process of identifying from among the multiple registered texts a registered text that is similar to the input text, and a second process of searching among the multiple insertion phrases for an insertion phrase corresponding to the input text, and the generating of the distribution information includes, in a case where an insertion phrase corresponding to the input text is identified at the second process, generating distribution information that indicates the related information corresponding to a text obtained by inserting the insertion phrase identified at the second process into the insertion section of the registered text identified at the first process, whereas generating distribution information that indicates the related information corresponding to a text corresponding to the registered text from which the insertion section has been removed in a case where an insertion phrase corresponding to the input text is not identified at the second process. An information management method according to another aspect of the present invention is an information management method for generating related information related to a guidance voice, and includes: identifying from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and generating the related information corresponding to a modified text that is partially different from the identified registered text.
A guide person UB managing the vehicle 200, such as a driver or a conductor, is the source of a voice G which provides guidance about transportation facilities (hereafter, simply referred to as "guidance voice") at an appropriate time. By the guidance voice G, voices on a variety of information may be provided regarding operations of a transportation facility including, for example, voices of announcements of a name of a next stop (station or bus stop) or transfer to other lines; voices of announcements concerning facilities located in the vicinity of the next stop (for example, tourist information); voices of announcements on an operation status of a transportation facility (for example, an occurrence of a stop signal, delay, or the like); voices of announcements cautioning care to be taken while on board, or care to be taken in boarding or in getting off transport; and voices of announcements upon occurrence of emergency. The information management system 100 of the first embodiment generates distribution information Q according to the guidance voice G spoken by the guide person UB, and transmits the information to the terminal device 30. The distribution information Q is information required for the terminal device 30 to present information related to the guidance voice G (hereafter referred to as "related information") to the user UA. The terminal device 30 of the first embodiment presents a text expressing the spoken content of the guidance voice G as the related information to the user UA. Therefore, it is possible, for example, for a hearing-impaired person who has difficulty in hearing the guidance voice G, to understand the content of the guidance voice G. The audio signal SG generated by the sound receiving device 14 is supplied as an audio signal SA to the sound outputting device 18 via the audio device 16. The audio device 16 executes audio processes, such as an amplification process and an adjustment process (for example, adjustment of frequency characteristics) for the audio signal SG, thereby generating the audio signal SA. The sound outputting device 18 is audio equipment (speaker) for outputting a sound corresponding to the audio signal SA supplied by the audio device 16. Specifically, a guidance voice G represented by the audio signal SG is outputted from the sound outputting device 18, for transmission to the user UA. For descriptive purposes, illustration of the D/A converter for converting the digital audio signal SA to analog format is omitted in the drawing. As will be understood from the above description, the voice guidance system 10 of the first embodiment is an audio system in which the distribution terminal 12 is connected with an existing in-car announcement system for outputting the guidance voice G from the sound outputting device 18 after processing by the audio device 16; the guidance voice G to be processed is received by the sound receiving device 14. It is of note that the configuration of the voice guidance system 10 is freely selected; for example, the elements of the distribution terminal 12, the sound receiving device 14, the audio device 16, and the sound outputting device 18 may be provided in a single apparatus. The audio signal SG generated by the sound receiving device 14 diverges from the path between the sound receiving device 14 and the audio device 16, and is supplied to the distribution terminal 12. Specifically, the audio signal SG is supplied to the distribution terminal 12 via a wired or wireless path. The distribution terminal 12 is an information device for providing the terminal device 30 with distribution information Q corresponding to the guidance voice G represented by the audio signal SG supplied from the sound receiving device 14. The distribution terminal 12 is realized by a portable terminal device, for example, a mobile phone, a smartphone, a tablet terminal, etc. The distribution terminal 12 of the first embodiment includes a control device 122 and a communication device 124, as illustrated in The control device 122 is a processing device (for example, a CPU (Central Processing Unit)) for controlling overall operation of the distribution terminal 12. Multiple functions for acquiring and distributing distribution information Q corresponding to the guidance voice G (a voice acquirer 52 and a signal processor 54) can be achieved by executing a program by the control device 122, the program being stored in a known recording medium (not shown), such as a magnetic recording medium or a semiconductor recording medium. The voice acquirer 52 acquires the audio signal SG representative of the guidance voice G from the sound receiving device 14. The audio signal SG acquired by the voice acquirer 52 is transmitted from the communication device 124 via the communication network 300 to the management apparatus 20. The management apparatus 20 receives the audio signal SG transmitted from the voice guidance system 10, and generates distribution information Q for instructing the terminal device 30 to present the related information related to the guidance voice G represented by the audio signal SG. The distribution information Q generated by the management apparatus 20 is transmitted from the management apparatus 20 to the voice guidance system 10. The communication device 124 receives the distribution information Q transmitted by the management apparatus 20 from the communication network 300. The signal processor 54 generates an audio signal SQ containing the distribution information Q received at the communication device 124 from the management apparatus 20 as a sound component. To generate the audio signal SQ by the signal processor 54, a known technique may be freely adopted. For example, a configuration in which a carrier wave, such as a sine wave having a predetermined frequency, is frequency-modulated with the use of the distribution information Q, thereby generating an audio signal SQ; or preferable is a configuration that sequentially executes spreading modulation of the distribution information Q with the use of a spreading code and frequency conversion with the use of a carrier wave of a predetermined frequency, thereby generating an audio signal SQ. The frequency band of the audio signal SQ is a frequency band within which sound output by the sound outputting device 18 and sound reception by the audio device 16 is possible. The frequency band of the audio signal SQ falls within a range of a frequency band (for example, from 18 kHz to 20 kHz) that is higher than a sound frequency band of sound such as a voice (for example, the guidance voice G), music, etc., audible to a user in an ordinary environment. However, the frequency band of the audio signal SQ may be freely set: for example, an audio signal SQ within the audible frequency band may be generated. The audio signal SQ generated by the signal processor 54 is supplied to the sound outputting device 18 as an audio signal SA after processing by the audio device 16. In other words, the audio signal SA including the sound component corresponding to the distribution information Q (audio signal SQ) is supplied to the sound outputting device 18, and the sound component corresponding to the distribution information Q is outputted as sound from the sound outputting device 18. The audio device 16 may combine the audio signal SG and the audio signal SQ to generate the audio signal SA. As will be understood from the above description, the sound outputting device 18 of the first embodiment serves as means for transmitting distribution information Q to the terminal device 30 (transmitter) via sound communication using sound (sound waves) that is aerial vibration acting as a transmission medium. In other words, the sound outputting device 18 outputs the guidance voice G received by the sound receiving device 14, and further transmits the distribution information Q to the terminal device 30 by output of sound including distribution information Q. In the above configuration, the sound outputting device 18 outputting the guidance voice G is also used for the transmission of the distribution information Q, and as a result the above configuration has an advantage in that the configuration of the voice guidance system 10 can be simplified in comparison with a configuration in which another device that is different from the sound outputting device 18 transmits the distribution information Q to the terminal device 30. The management apparatus 20 shown in The control device 22 is a processing device (for example, a CPU) that controls overall operation of the management apparatus 20. The communication device 26 communicates with the distribution terminal 12 via the communication network 300. For example, the communication device 26 receives the audio signal SG transmitted from the distribution terminal 12, and transmits distribution information Q corresponding to the audio signal SG to the distribution terminal 12. The storage device 24 stores programs executed by the control device 22 and various data used by the control device 22. A known recording medium, such as a magnetic recording medium or a semiconductor recording medium, or a combination of multiple types of recording media may be freely adopted as the storage device 24. The storage device 24 of the first embodiment stores a guidance table TA. As shown in The control device 22 in The voice analyzer 62 identifies a text (hereafter referred to as "input text") L representative of the speech content of the guidance voice G by use of speech recognition performed on the audio signal SG received by the communication device 26 from the distribution terminal 12. In other words, the input text L is a text spoken by the guide person UB. For performing speech recognition to represent the guidance voice G there may be freely adopted a known technique, for example, recognition processing utilizing an audio model, such as the HMM (Hidden Markov Model), and a language model indicating language constraints. As mentioned above, the guide person UB basically speaks one of the texts X registered in the announcement book, etc., prepared beforehand. Accordingly, ideally, the input text L identified through the speech recognition of the guidance voice G by means of the speech analyzer 62 matches any one of the registered texts X registered in the guidance table TA. Actually, however, a recognition error may occur in the speech recognition by the voice analyzer 62 because of the pronunciation traits (habits) unique to each individual guide person UB, noise and the like around the sound receiving device 14, etc. Therefore, the input text L and the registered text X may be similar to each other, but be different partially from each other. In view of the above, the text identifier 64 of the first embodiment identifies a registered text X similar to the input text L identified by the voice analyzer 62 among the multiple different registered texts X. Specifically, the text identifier 64 identifies a registered text X similar to the input text L identified by the speech analyzer 62 among the multiple registered texts X registered in the guidance table TA, and identifies an insertion phrase Y corresponding to the input text L among the multiple insertion phrases Y corresponding to the registered text X. As shown in The second process S52 shown in The information generator 66 in As shown in On the other hand, if an insertion phrase Y is not identified at S52 (if the decision at S61 is negative), the information generator 66 generates distribution information Q that indicates a modified text Z corresponding to the registered text X identified at the first process S51 (that is, a text which is partially different from the registered text X) as related information (S63). Specifically, the information generator 66 obtains from the guidance table TA identification information DZ of the modified text Z corresponding to the registered text X, and generates the distribution information Q containing the identification information DZ. Specific phrases other than the multiple insertion phrases Y (hereafter referred to as "specific phrases") may be registered in the guidance table TA in advance, and each of the specific phrases may be compared with the input text L at the second process S52, as being similar to each of the multiple insertion phrases Y, so as to decide whether the specific phrase is included in the input text L. For example, phrases that are highly likely to be pronounced incorrectly by the guide person UB, or phrases that may be misrecognized by the voice analyzer 62 are selected in advance as the specific phrases. In a case where the input text L includes one of the specific phrases, the information generator 66 decides that an insertion phrase Y is not identified at the second process S52 (the decision at S61 is negative). In other words, "insertion phrase Y is not identified" is intended to include a case where a specific phrase other than the insertion phrase Y is found in the input text L, in addition to the above example in which an insertion phrase Y is not actually identified. Upon generating the distribution information Q at the process (S62 or S63) shown above, the information generator 66 transmits the distribution information Q from the communication device 26 to the distribution terminal 12 of the voice guidance system 10 (S7). As mentioned above, in the audio guidance system 10, the signal processor 54 and the audio device 16 generate an audio signal SA containing the distribution information Q received from the management apparatus 20 as the sound component, and the sound outputting device 18 outputs a sound corresponding to the audio signal SA (that is, the sound including the distribution information Q). Practically, at a stage where the guide person UB finishes voicing the guidance voice G, the input text L is identified by the voice analyzer 62, and generation and transmission of the distribution information Q are executed. Accordingly, the sound of the distribution information Q is outputted from the sound outputting device 18 at a time point behind the sound output of the guidance voice G. The storage device 36 stores programs executed by the control device 34 and various data used by the control device 34. The control device 34 is a processing device (for example, a CPU) that controls overall operation of the terminal device 30. The control device 34 of the first embodiment executes a program stored in the storage device 36, thereby realizing multiple functions (information extractor 72 and presentation controller 74) for presenting to the user UA related information according to distribution information Q. The information extractor 72 extracts the distribution information Q with demodulation of the audio signal SB generated by the sound receiving device 32. Specifically, the information extractor 72 performs a filtering process for emphasizing band components within the frequency band including the sound component of the distribution information Q and a demodulation process corresponding to the modulation process in the signal processor 54 for the audio signal SB to extract the distribution information Q. The presentation controller 74 causes the presentation device 3 8 to present related information R indicated by the distribution information Q extracted by the information extractor 72. The presentation device 38 presents the related information R indicated by the presentation controller 74 to the user UA. The presentation device 38 of the first embodiment is a display device (for example, a liquid crystal display panel) for displaying the related information R. The guidance table TB stored in the storage device 36 is used for the process in which the presentation controller 74 identifies the related information R indicated by the distribution information Q. As shown in If the distribution information Q contains identification information DX and identification information DY (that is, if a text resulting from insertion of an insertion phrase Y into a registered text X as a result of identifying the insertion phrase Y at the second process S52 is indicated as the related information R), the presentation controller 74 identifies the related information R of the identification information DR corresponding to the combination of the identification information DX and the identification information DY in the guidance table TB, and causes the presentation device 38 to present it. Therefore, a text obtained by inserting the insertion phrase Y included in the input text L into the insertion section B of the registered text X that is similar to the speech content of the guidance voice G (the input text L) (that is, one or more sentences generally coincident with the speech content of the guidance voice G) is presented to the user UA as related information R. On the other hand, if the distribution information Q contains identification information DZ (that is, if an insertion phrase Y is not identified at the second process S52, and therefore a modified text Z is indicated as related information R), the presentation controller 74 identifies the related information R of the identification information DR corresponding to the identification information DZ in the guidance table TB, and causes the presentation device 38 to present the information. Therefore, a modified text Z (that is, one or more sentences that are partially different from the speech content of the guidance voice G) that is partially changed from the registered text X similar to the speech content of the guidance voice G is presented to the user UA as related information R. As will be understood from the above description, the registered texts X may be defined as texts that are used for comparison with the input text L for presentation to the user UA, whereas the modified texts Z may be defined as texts that are used for presentation to the user UA, and are not used for comparison with the input text L. In the above example, the combination of the identification information DX and the identification information DY is shown as the identification information DR for the related information R, but each of the identification information DX and the identification information DY may be registered as identification information DR for related information pieces R (the registered text X and the insertion phrase Y) with the guidance table TB. The presentation controller 74 may acquire related information R (a registered text X) having the identification information DR corresponding to the identification information DX specified by the distribution information Q, acquire related information R (insertion phrase Y) having the identification information DR corresponding to the identification information DY specified by the distribution information Q, and cause the presentation device 38 to present a text obtained by combining related information R (a registered text X) and related information R (insertion phrase Y) to the user UA as the related information R. Upon receiving the audio signal SG at the communication device 26, the management apparatus 20 sequentially executes the identification of the input text L by the voice analyzer 62 (S4), the identification processing by the text identifier 64 (S5: S51, S52), the generation of the distribution information Q by the information generator 66 (S6: S61 to S63), and the transmission of the distribution information Q (S7). At the distribution terminal 12, an audio signal SQ including the sound component of the distribution information Q is generated (Sg), and the distribution information Q is transmitted to the terminal device 30 as a result of reproduction of sound by the sound outputting device 18 on the basis of the audio signal SQ (S9). The sound outputted by the sound outputting device 18 is received by the sound receiving device 32 of the terminal device 30 (S10). In the terminal device 30, the information extractor 72 extracts the distribution information Q from the audio signal SB generated by the sound receiving device 32 (S11) by receiving sound, and the information generator 66 acquires related information R corresponding to the distribution information Q from the guidance table TB, and causes the presentation device 38 to present the information to the user UA (S12). Therefore, while listening to the guidance voice G outputted from the sound outputting device 18, the user UA can confirm the related information R corresponding to the guidance voice G by way of the display of the presentation device 38. As will be understood from the above description, in the first embodiment, a registered text X similar to the input text L identified by speech recognition of the guidance voice G is identified from among the multiple registered texts X. Therefore, as compared with, for example, a configuration in which the input text L identified from the guidance voice G is presented as related information R to the user UA of the terminal device 30, more suitable related information R can be presented to the user UA with less influence being caused by a voice recognition error. In addition, since a modified text Z partially different from the registered text X identified by the text identifier 64 is presented to the user UA of the terminal device 30 as related information R, an advantage is obtained in that the modified text Z that is partially changed from the registered text X is presented to the user UA of the terminal device 30 while the registered texts X that are supposed to be used as guidance voices G are used for comparison with the input text L. In the first embodiment, if an insertion phrase Y included in the input text L is identified at the second process S52, a text obtained by inserting the insertion phrase Y into the insertion section B of the registered text X is presented as related information R, whereas if an insertion phrase Y is not identified at the second process S52, a modified text Z is presented as related information R. Accordingly, even if the guide person UB makes a mispronunciation (for example, if the guide person UB speaks a phrase other than suitable phrases) or the voice analyzer 62 makes a recognition error, it is possible to reduce a possibility of presenting to the user UA related information R containing an incorrect phrase. A second embodiment of the present invention will now be described. In the following description, the same reference symbols are used for identifying elements of which actions or functions are the same as in the first embodiment, and detailed explanation of the elements will be omitted as appropriate. In the guidance table TA of the second embodiment, for each registered text X, multiple insertion phrases Y that can be inserted into the insertion section B of the registered text X are registered. For example, for the aforementioned registered text X1, multiple insertion phrases Y (Y11, Y12, Y13, ...), such as "because of vehicle inspection", "because of signal failure", and "because of entry of a person in the railway" are registered. For the registered text X2, "We will soon make a stop [ ]. The doors on the left side will open.", multiple insertion phrases Y (Y21, Y22, Y23, ...), such as "at Tokyo station", "at Akihabara station", and "at Ueno station", are registered. In the second embodiment, the modified text Z is not registered with the guidance table TA. As similar to the first embodiment, the text identifier 64 of the second embodiment executes a first process S51 of identifying a registered text X similar to the input text L from among the multiple registered texts X, and a second process S52 of searching among the multiple insertion phrases Y for an insertion phrase Y corresponding to the input text L corresponding to the registered text X. Similar to the first embodiment, the information generator 66 decides, at S61, whether an insertion phrase Y corresponding to the input text L is identified at the second process S52. If an insertion phrase Y is identified (if the decision at S61 is affirmative), the information generator 66 generates distribution information Q that indicates a combination of the registered text X and the insertion phrase Y (S62). On the other hand, if an insertion phrase Y is not identified at the second process S52 (if the decision at S61 is negative), the information generator 66 generates distribution information Q that indicates the registered text X as related information R (specifically, distribution information Q including the identification DX for the registered text X) (S63), and transmits the related information R from the communication device 26 to the voice guidance system 10 (S7). As similar to the first embodiment, sound including the distribution information Q is outputted from the sound outputting device 18, and the distribution information Q is extracted from the audio signal SB at the terminal device 30. If an insertion phrase Y is identified at the second process S52, as similar to the first embodiment, related information R corresponding to a combination of the identification information DX and the identification information DY is presented to the user UA by the presentation device 38. On the other hand, if an insertion phrase Y is not identified at S52, a registration text X corresponding to the identification information DX designated by the distribution information Q (preferably, a text obtained by removing the insertion section B of the registered text X) is presented to the user UA as related information R. In other words, if an insertion phrase Y is not identified at the second process S52, the information generator 66 of the second embodiment generates related information R that indicates a text obtained by removing the insertion section B from the registered text X as related information R. If there is announced an unusual reason for stopping, such as a case where there is announced "We have stopped due to a meteorite impact. We apologize for the delay. Please wait for resumption,", there is presented, to the user UA, linguistically natural related information R that is "We have stopped. We apologize for the delay. Please wait for resumption." In addition, where there is identified an erroneous input text L as a result of a recognition error by the voice analyzer 62, for example when a guidance voice G announces "We have made a stop because of a malfunction of the vehicle. We apologize for the delay. Please wait for resumption", an erroneous input text L reads "We have made a stop because of trouble on the company president. We apologize for the delay. Please wait for resumption" (It is of note here that in Japanese "malfunction of the vehicle" is "sharyou koshou", whereas "trouble on the company president" is "syachou koshou"), even in this case, linguistically natural related information R that is "We have made a stop. We apologize for the delay. Please wait for resumption." is presented to the user UA. In the second embodiment as well as the first embodiment, since a registered text X similar to the input text L identified by the speech recognition of the guidance voice G is identified from among the multiple registered texts X, suitable related information R can be presented to the user UA with less influence being caused by speech recognition error. In addition, in the second embodiment, if an insertion phrase Y corresponding to the input text L is identified at the second process S52, distribution information Q that indicates a text obtained by inserting the insertion phrase Y into the insertion section B of the registered text X as related information R is generated, whereas if an insertion phrase Y corresponding to the input text L is not identified at the second process S52, distribution information Q that indicates a text obtained by removing the insertion section B from the registered text X as related information R is generated. Accordingly, even if the guide person UB makes a mispronunciation (for example, if the guide person UB speaks a phrase other than predicted insertion phrases Y) or a recognition error is made for the guidance voice G, it is possible to reduce a possibility of presenting to the user UA related information R containing an incorrect phrase. In the third embodiment, it is assumed that the information management system 100 is used to provide information to the user UA located in a commercial facility (for example, a shopping mall). The voice guidance system 10 of the information management system 100 is provided in a commercial facility, whereas the management apparatus 20 is connected to the communication network 300 as similar to the first embodiment. As shown in The text identifier 64 of the third embodiment identifies a registered text X similar to the input text L from among the multiple registered texts X in the guidance table TA (SA1). Processing similar to the first process S51 shown in the first embodiment is used for the identification (SA1) of the registered text X. As described above, in the third embodiment, since a registered text X similar to the input text L identified by the speech recognition of the guidance voice G is identified, suitable related information R can be presented to the user UA with a recognition error in speech recognition having little influence, as in the first embodiment. The information generator 66 generates distribution information Q that indicates as related information R a modified text Z corresponding to the registered text X identified by the text identifier 64 (SA2). Specifically, the information generator 66 generates distribution information Q including the identification information DZ of the modified text Z associated with the registered text X in the guidance table TA. The information generator 66 transmits the distribution information Q generated by the above procedure from the communication device 26 to the voice guidance system 10 (SA3). The subsequent processing is the same as the first embodiment. In other words, sound including distribution information Q is outputted from the sound outputting device 18. At the terminal device 30, the presentation device 38 presents the modified text Z indicated by the distribution information Q extracted from the audio signal SB as related information R to the user UA. Therefore, for example, in conjunction with a guidance voice G, "XYZ from ABC city is waiting for you. Please meet your party at the information booth.", a modified text Z1, reading "Your companion is waiting for you. Anyone who may know of this person, Please come to the information desk." is presented to the user UA by the presentation device 38. In addition, in parallel to a guidance voice G, "The owner of a red van in the parking lot, registration number 'Ward A 12-3456', the headlights are left on. Please return to your car.", a modified text Z2, reading "The owner of a red van in the parking lot, the headlights are left on. Please return to your car." is presented to the user UA by the presentation device 38. In other words, the modified text Z that excludes personal information (the place of residence, name, registration number, etc.) from the guidance voice G is presented by the presentation device 3 8 to the user UA. Therefore, it is possible to protect personal information. In the third embodiment, similar to the first embodiment, since a modified text Z partially different from the registered text X identified by the text identifier 64 is presented to the user UA of the terminal device 30 as related information R, an advantage is obtained in that the modified text Z, partially changed from the registered text X, is presented to the user UA of the terminal device 30 while the registered text X prepared to be voiced as guidance voices G is used for comparison with the input text L. In the first to third embodiments shown above, related information R in the same language as in the guidance voice G is presented to the user UA, but a text translated from the guidance voice G to another language can be presented as related information R to the user UA of the terminal device 30. Specifically, texts in languages different from that of guidance voices G may be registered with the guidance table TB as related information R. According to the above configuration, since related information R corresponding to a translation text of a guidance voice G is presented to the user UA in parallel to sound reproduction of the guidance voice G, the related information will be useful for foreigners who cannot understand the language of the guidance voice G. As will be understood from the above examples, related information R in the first and third embodiments can be defined as information corresponding to the modified texts Z partially different from the registered texts X identified by the text identifier 64, and include, in addition to the modified texts Z themselves, voice expressions of the modified texts Z, translated modified texts, and voice expressions of the translated modified texts. Related information R in the second embodiment can be defined as information corresponding to the texts obtained by inserting the insertion phrases Y into the insertion section B of the registered texts X (see S62), or as information pieces corresponding the texts obtained by removing the insertion section B from the registered texts X (see S63), and include, in addition to the texts themselves, voice expressions of the texts, translated texts, and voice expressions of the translated texts. In the first to third embodiments, the distribution information Q that instructs the terminal device 30 to present related information R is sent from the information management system 100. In the fourth embodiment, the information management system 100 generates related information R corresponding to the guidance voice G, and provides it to the user UA. Operation of the voice analyzer 62 and the text identifier 64 is the same as in the above-described embodiments. Thus, similarly to the above-described embodiments, it is possible to present appropriate related information R to the user UA with reduced influence of a speech recognition error. If an insertion phrase Y is identified at the second process S52 (if the decision at S61 is affirmative), the information generator 66 of the fourth embodiment generates, as related information R, a text that is translated to another language from a text having the registered text X into which the insertion phrase Y is inserted. On the other hand, if an insertion phrase Y is not identified at the second process S52 (if the decision at S61 is negative), the information generator 66 generates, as related information R, a text that is translated to another language from the modified text Z corresponding to the registered text X identified at the first process S51. In other words, the information generator 66 in the fourth embodiment generates related information piece R corresponding to a modified text Z partially changed from the registered text X identified by the text identifier 64. For translation of texts, for example, a known machine translation technique can be adopted in a freely-selected manner. Related information R generated by the information generator 66 is transmitted to the distribution terminal 12 of the voice guidance system 10. The signal processor 54 of the distribution terminal 12 generates an audio signal SQ by means of speech synthesis to which related information R is applied. The audio signal SQ in the fourth embodiment is a signal representing a sound of a spoken text specified by related information R. To generate the audio signal SQ, known speech synthesis can be freely adopted. The audio signal SQ generated by the signal processor 54 is supplied to the sound outputting device 18 via the audio device 16. Thus, a speech sound of the text identified by related information R is outputted from the sound outputting device 18. In other words, following the guidance voice G spoken by the guide person UB, another voice translated from the guidance voice G to another language is outputted from the sound outputting device 18 to the user UA. For example, if a translation text of a modified text Z is generated as related information R, a speech sound of the translation text of the modified text Z partially changed from the registered text X corresponding to the guidance voice G is outputted following the guidance voice G. In the above description, a configuration based on the first embodiment is discussed, but identification of a text in the second embodiment or the third embodiment may be applied to the fourth embodiment. For example, if an insertion phrase Y is identified at the second process S52 (if the decision at S61 is affirmative), the information generator 66 in the fourth embodiment generates, as related information R, a text that is translated to another language from a text having the registered text X into which the insertion phrase Y is inserted. On the other hand, if an insertion phrase Y is not identified at the second process S52 (if the decision at S61 is negative), the information generator 66 generates, as related information R, a text that is translated to another language from a text obtained by removing the insertion section B from the registered text X identified at the first process S51. Therefore, the voice of the translation text of the registered text X excluding the insertion section B is outputted from the sound outputting device 18 following the guidance voice G. In addition, assuming the third embodiment, the information generator 66 may generate related information R representing a text obtained by translating the modified text Z corresponding to the registered text X identified by the text identifier 64 into another language. With this configuration, the voice of the translation text of the modified text Z partially changed from the registered text X corresponding to the guidance voice G is outputted in conjunction with the guidance voice G. As will be understood from the above description, the information management system 100 in the fourth embodiment is a system that generates related information R related to guidance voices G (and provides the user UA with related information R), and includes the text identifier 64 that identifies a registered text X similar to the input text L identified by speech recognition of the guidance voice G from among the multiple registered texts X, and the information generator 66 that generates related information R corresponding to the registered text X identified by the text identifier 64. A typical example of related information R corresponding to the registered text X is a translation text of the modified text Z that is partially different from the registered text X, or a translation text of a text resulting from deletion of the insertion section B of the registered text X. In the above description, a configuration for outputting a speech sound of a text indicated by related information R from the sound outputting device 18 is shown; however, the method for outputting related information R is not limited to the above example. For example, it is also possible to display the text indicated by related information R on the display device. The above-shown embodiments may be variously modified. Specific modifications are shown below. Two or more variations freely selected from the following variations may be appropriately combined unless they are contradictory to each other.
An information management system according to a preferred mode (Mode 1) of the present invention is an information management system for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device: a text identifier configured to identify from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and an information generator configured to generate distribution information that indicates to the terminal device the related information corresponding to a modified text that is partially different from the registered text identified by the text identifier. In Mode 1, a registered text similar to the input text representative of the guidance voice is identified from among the multiple registered texts. Therefore, as compared with, for example, a configuration in which an input text identified by speech recognition analysis for the guidance voice or an input character entered by a guide person is presented as related information to the user of the terminal device, more suitable related information can be presented to the user. In addition, distribution information that indicates, to the terminal device, related information corresponding to the modified text that is partially different from the registered text identified by the text identifier is generated. Accordingly, an advantage is obtained in that related information that is partially different from the registered text is presented to the user of the terminal device while the registered texts intended for use as guidance voices are used for comparison with the input text. For example, it is possible to present, to the user, related information obtained by removing items not suitable for presenting from the terminal device to the user (for example, personal information) from the guidance voice or appropriate related information obtained by modifying a part that cannot readily be processed with speech recognition. The first embodiment and the third embodiment described above correspond to examples of Mode 1. In a preferred example (Mode 2) of Mode 1, each of the registered texts includes an insertion section in which a selected one of multiple insertion phrases is inserted, with the text identifier being configured to execute a first process of identifying a registered text from among the multiple registered texts that is similar to the input text, and a second process of searching among the multiple insertion phrases for an insertion phrase corresponding to the input text for the registered text, and the information generator being configured to, in a case where an insertion phrase corresponding to the input text is identified at the second process, generate distribution information that indicates the related information corresponding to a text obtained by inserting the insertion phrase identified at the second process into the insertion section of the registered text identified at the first process, whereas to generate distribution information that indicates the related information corresponding to a modified text that is partially different from the registered text identified at the first process where an insertion phrase corresponding to the input text is not identified at the second process. In Mode 2, if an insertion phrase corresponding to the input text is identified at the second process, distribution information that indicates related information corresponding to a text obtained by inserting the insertion phrase into the insertion section of the registered text is generated, whereas if an insertion phrase corresponding to the input text is not identified at the second process, distribution information that indicates related information corresponding to a modified text that is partially different from the registered text is generated. Accordingly, even if the guide person makes a mispronunciation (for example, if the guide person speaks a phrase other than predicted insertion phrases) or a recognition error occurs for the guidance voice, it is possible to reduce a possibility of presenting to the user related information containing an inappropriate phrase. The first embodiment described above corresponds to an example of Mode 2. In a preferred example (Mode 3) of Mode 1, the information generator is configured to generate distribution information that indicates the related information corresponding to the modified text obtained by deleting part of the registered text identified by the text identifier. In Mode 3, distribution information is generated that indicates related information corresponding to the modified text obtained by deleting part of the registered text. Accordingly, for example, it is possible to present to the user related information obtained by deleting from the guidance voice information not suitable for presentation from the terminal device to the user (for example, personal information). The third embodiment corresponds to an example of Mode 3. An information management system according to a preferred mode (Mode 4) of the present invention is an information management system information management system for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, and includes: a text identifier configured to identify from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and an information generator configured to generate the distribution information that indicates to the terminal device the registered text identified by the text identifier, each of the registered texts including an insertion section for insertion of a selected one of multiple insertion phrases, with the text identifier being configured to execute a first process of identifying a registered text that is similar to the input text from among the multiple registered texts, and a second process of searching among the multiple insertion phrases for an insertion phrase corresponding to the input text, and the information generator being configured to, in a case where an insertion phrase corresponding to the input text is identified at the second process, generate distribution information that indicates the related information corresponding to a text obtained by inserting the insertion phrase identified at the second process into the insertion section of the registered text identified at the first process, whereas to generate distribution information that indicates the related information corresponding to a text corresponding to the registered text from which the insertion section has been removed in a case where an insertion phrase corresponding to the input text is not identified at the second process. In Mode 4, a registered text similar to the input text representative of the guidance voice is identified from among the multiple registered texts. Therefore, as compared with, for example, a configuration in which an input text identified by speech recognition analysis for the guidance voice or an input character entered by a guide person is presented as related information to the user of the terminal device, more suitable related information can be presented to the user. In addition, if an insertion phrase corresponding to the input text is identified at the second process, distribution information that indicates related information corresponding to a text obtained by inserting the insertion phrase into the insertion section of the registered text is generated, whereas if an insertion phrase corresponding to the input text is not identified at the second process, distribution information that indicates related information corresponding to a text corresponding to the registered text from which the insertion section has been removed is generated. Accordingly, even if the guide person makes mispronunciation (for example, if the guide person speaks a phrase other than predicted insertion phrases) or a recognition error is made for the guidance voice, it is possible to reduce likelihood of presenting to the user related information containing an incorrect phrase. The second embodiment corresponds to an example of In a preferred example (Mode 5) of any one of Modes 1 to 4, the information management system further includes a sound outputter configured to output the guidance voice and to output a sound including the distribution information to transmit the distribution information to the terminal device. In Mode 5, the sound outputter that outputs a guidance voice is also used for sound output of the distribution information (that is, sound communication with the use of sound with aerial vibration acting as a transmission medium). Therefore, it is possible to simplify the configuration of the information management system compared with a configuration in which the distribution information is transmitted to the terminal device by means of a device that is different from the sound outputter used for sound output of the guidance voice. An information management system according to a preferred mode (Mode 6) of the present invention is an information management system for generating related information related to a guidance voice, and includes: a text identifier configured to identify from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and an information generator configured to generate the related information corresponding to a modified text that is partially different from the registered text identified by the text identifier. In Mode 6, there is identified from among the multiple registered texts a registered text similar to the input text representative of the guidance voice. Therefore, as compared with, for example, a configuration in which an input text identified by speech recognition of the guidance voice or an input character entered by a guide person is presented as related information to the user of the terminal device, more suitable related information can be presented to the user. In addition, related information (for example, a translation text of the modified text) corresponding to the modified text that is partially different from the registered text identified by the text identifier is generated. Accordingly, there is an advantage that related information that is partially different from the registered text is presented to the user, while the registered texts supposed to be used as guidance voices are used for comparison with the input text. The third embodiment corresponds to an example of Mode 6. In a preferred example (Mode 7) of any of Modes 1 to 6, the text identifier is configured to identify a registered text that is similar to the input text identified by speech recognition of the guidance voice from among the multiple registered texts. In Mode 7, since the input text is identified by speech recognition of the guidance voice, there is an advantage that the guide person does not need to manually input the input text. An information management method according to a preferred mode (Mode 8) of the present invention is an information management method for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, and includes: identifying from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and generating distribution information that indicates to the terminal device the related information corresponding to a modified text that is partially different from the identified registered text. According to Mode 8, there is achieved the same effect as in the information management system according to Mode 1. An information management method according to a preferred mode (Mode 9) of the present invention is an information management method for generating distribution information for indicating to a terminal device related information related to a guidance voice for presentation of the related information to a user by the terminal device, and includes: identifying from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and generating distribution information that indicates to the terminal device the identified registered text, each of the registered texts including an insertion section in which a selected one of multiple insertion phrases is inserted, and the identifying of a registered text includes executing a first process of identifying from among the multiple registered texts a registered text that is similar to the input text, and a second process of searching among the multiple insertion phrases for an insertion phrase corresponding to the input text. The generating of the distribution information includes, in a case where an insertion phrase corresponding to the input text is identified at the second process, generating distribution information that indicates the related information corresponding to a text obtained by inserting the insertion phrase identified at the second process into the insertion section of the registered text identified at the first process, whereas generating distribution information that indicates the related information corresponding to a text corresponding to the registered text from which the insertion section has been removed in a case where an insertion phrase corresponding to the input text is not identified at the second process. According to Mode 9, the same effects as in the information management system according to Mode 4 is achieved. An information management method according to a preferred mode (Mode 10) of the present invention is an information management method for generating related information related to a guidance voice, and includes: identifying from among multiple different registered texts a registered text that is similar to an input text representative of the guidance voice; and generating the related information corresponding to a modified text that is partially different from the identified registered text. According to the Mode 10, the same effects as in the information management system according to Mode 6 are achieved.
|