WO2000078024A1 - System and method for voice-controlled telecommunications services - Google Patents
System and method for voice-controlled telecommunications services Download PDFInfo
- Publication number
- WO2000078024A1 WO2000078024A1 PCT/SE2000/001240 SE0001240W WO0078024A1 WO 2000078024 A1 WO2000078024 A1 WO 2000078024A1 SE 0001240 W SE0001240 W SE 0001240W WO 0078024 A1 WO0078024 A1 WO 0078024A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- customer
- service
- voice
- updating
- records
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/487—Arrangements for providing information services, e.g. recorded voice services or time announcements
- H04M3/493—Interactive information services, e.g. directory enquiries ; Arrangements therefor, e.g. interactive voice response [IVR] systems or voice portals
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M2201/00—Electronic components, circuits, software, systems or apparatus used in telephone systems
- H04M2201/40—Electronic components, circuits, software, systems or apparatus used in telephone systems using speech recognition
Definitions
- the present invention relates to a method and device at a telecommunications system that makes possible to automatically update functions on a general level from the user' s interception computer to collection of the information the application need, to be able to handle the user' s catalogue for voice-controlled telecommunications services .
- Telephone services where a calling party can be forwarded to extensions, give orders for articles/services, provide or obtain information, etc, without the need to talk to a telephone operator are today very frequent. Such services are generally based on that a number of preselections have been programmed by the customer of the service, which pre-selections are presented to the calling party by a recorded voice. These pre-selections then are accessible to the calling party by keying on his/her keyset code telephone. Navigation in a pre-selection programmed exchange service, or, for instance, in a service for ordering of articles, often can be made in several steps with repeated selections by means of keying.
- a lot of voice-controlled telecommunications services today require information from a plurality of different sources.
- an operator who offers the voice-controlled service need information about the customer, for instance derived from an external database belonging to the customer.
- necessary information for speech recognition at the operator's voice service device i.e. a kind of dictionary.
- speech recognition at the operator's voice service device
- information about which words or phrases the system shall react to and necessary sound files that shall be fed back to the user, i.e. the calling party.
- One problem today is that the work to collect this types of information to a common system is not done automatically, but separately from each other and with primarily manual handling.
- voice-controlled systems require a considerable database with words for recognition, and must be able to understand different pronunciations, to make it possible to utilise the service in an efficient way.
- the aim of the present invention is provide a voice control system that solves these problems.
- the invention solves the above mentioned problems by a system and a method to automatically derive necessary information for a voice-controlled service, including sound data.
- the system includes a general database for words that shall be possible to use in voice-controlled services and where each word is associated with sound data, including return value, transcriptions and sound files.
- An updating system is arranged to automatically derive information from an external database with words that shall be included in the service, and sound data from the word database. The information from the two databases is distributed automatically to the system for the voice-controlled service.
- a new database is after that automatically built up that corresponds to the external database, but which now also contains return values and sound files.
- the application also generates grammar with transmitted return values for words or combinations of words to which the service shall react.
- the invention consequently relates to an updating system for a voice-controlled service in a telecommunications system, which service is offered by an operator, for instance Telia, and which includes means to recognise and react to a voice command given by a calling party to a customer who subscribes to the service.
- the invention includes a customer data device for automatic obtaining of catalogue information from the customer to the operator, which customer data device is connected to the updating system. Further are included in the updating system a word database with sound data for a voice service, and an updating device which is connected to the customer data device and the word database, arranged to, in the word database, automatically search for and derive words defined in said catalogue information.
- the updating device is further arranged to automatically create and update files with information from the customer' s catalogue information with belonging sound data, which shall be used for the voice-controlled service.
- Figure 1 shows a block diagram of the automatic call routine according to the invention
- FIG. 2 shows units and flows in the voice control system according to the invention. Detailed description of preferred embodiments
- Telia's OrdBok (Telia's Dictionary) which is arranged to contain sound data for a voice-controlled service.
- Telia's OrdBok is built up as a general SQL-database for words that shall be possible to use in different voice-controlled services. It contains for each word the following sound data: Return value, transcriptions (descriptions of pronunciation) and sound files. All words have a return value, but there can be more than one word that has the same return value. In that way, for instance the words (names) Eriksson and Erixon, can have the same return value, as a suggestion with the spelling Eriksson. The words have different spelling, but are pronounced in the same way.
- return value is meant an index for the word in the database, and can also be called normalised spelling.
- the updating routine for the voice-controlled service automatically derives information from an external database and sound data from the word database for the words that shall be included in the service.
- Information from the two databases is automatically distributed to the system for the voice-controlled service.
- ap . ' cation for the service is then a new database automat-, -.ally built up that corresponds to the external database but which now also includes return values and sound files.
- the application also generates grammar with transmitted return values for words or combination of words, to which the service shall react.
- the application generates a dictionary for the different pronunciation (transcriptions) of the words in question. Words that cannot be found in Telia's OrdBok are automatically registered as "residual items/records" for manual handling with transcription and recording.
- the sound data of the words later have been entered, the information will be delivered to the service at coming updatings .
- Telia's OrdBok is a word database with necessary sound data for a voice-controlled service. It is worked out to be used together with the automatic updating routine AUR for delivery to Telia's voice control service Telia DigiTala, but due to its construction it is also of current interest for use by other voice controlled services. Telia's OrdBok consequently is database for the words in the application Telia DigiTala that require voice recognition and playback with sound file.
- the operator for instance Telia, which provides the voice control service
- the customer which subscribes to the voice control service, for instance a company that utilises the service as a telephone exchange, and 3) a user or a calling party who, by means of voice- commands, can utilise the voice control service for wanted purpose.
- a customer defines necessary data in a file on an interception computer HVD .
- This interception computer is normally located at the customer's, and is arranged to tra " ⁇ z HVD-files to the operator. This is done, prefer-..- iy, over Internet to an FTP-server at the operator ' s, which server is included in the voice control system.
- an FTP-server at the operator ' s, which server is included in the voice control system.
- Both these are preferably lower-case letters with underscore instead of space type of word (for instance first name, family name, department etc.) transcription of normal pronunciation transcription of alternative pronunciation (one or more) reference to sound file for normal pronunciation. Also references to sound files for alternative pronunciation can be included. voice (ordinary or reserve voice) status. In status is included date for AUR's obtaining of sound file to respective platform (PC or SRP) .
- PC or SRP platform
- Telia's OrdBok (dictionary) is arranged to:
- Telia's OrdBok the transcriptions are unique within each type of word for each return value, i.e. a transcription exists only one time in Telia's OrdBok for a type of word.
- the reason for the addition "within each type of word" depends on that a word can exist as different types of word, for instance Helen as first name and family name.
- Telia's OrdBok contains words with both transcription and sound file, and it is updated running. New words are transcribed and a studio attends to that new-recorded sound files are available in the updating system.
- Telia's OrdBok for Telia DigiTala correspond to the text that is in respective field in the customer's HVD-file as above.
- the words that are entered correspond to the text the customer has written in corresponding fields.
- Sven Ingvar and Sven- Ingvar are entered in Telia's OrdBok as one word.
- Petterson Lundstr ⁇ m and Petterson-Lundstr ⁇ m are entered as one word.
- Alternative spellings shall in Telia's OrdBok be handled so that each spelling is a new word.
- Telia DigiTala is a voice-controlled service which makes it possible for a user to automatically be connected to an e tension in an exchange by using the voice only.
- Catalogue information for the customer is derived from the customer' s interception computer and is updated for the service Telia DigiTala by a separate updating system according to the invention.
- the system consists primarily of an automatic updating system AUR (Automatic Updating Routine) .
- AUR is arranged to reuse recorded sound files and transcriptions collected in a separate database, preferably Telia's OrdBok.
- a separate database preferably Telia's OrdBok.
- the updating system provides both information for applications on PC-platform and SRP- platform.
- the updating system AUR for Telia DigiTala attends to that the customer's catalogue in Telia DigiTala is automatically updated according to the customer's wishes. In a preferred embodiment this is done at fixed points of time. Alternatively the updating can be made at request from the customer, or the operator.
- the basis of the updating is the customer's HVD and the customer's request about which information in HVD that shall be accessible in Telia DigiTala.
- the updated information is made accessible to the application on the platform, PC or SRP. Separately, outside the updating system, the information is distributed within the platforms.
- Customer data registered as records in an HVD-file are transmitted to the updating system, preferably from the customer' s interception computer to an FTP-server over Internet, as has previously been described, from where the updatir. ⁇ system derives the HVD-file.
- Telia's Ordbok If the word for a group membership does not exist in Telia's Ordbok, lacking word is automatically registered as residual item/record, but the personal record will yet be included in Telia DigiTala. In a preferred embodiment it then will be possible to call such a record, but if there are a plurality of persons with the same pronunciation of name, the application is arranged to connect to telephone operator until the word of the group membership has been entered in Telia's Ordbok. For all records there consequently shall be a group membership to make it possible to distinguish between persons with the same pronunciation of name. It can, for instance, be division, department or place, which is determined in connection with the ordering of the service.
- the new updating for new records is primarily deliver d by an "ordinary" voice, i.e. a voice which is used for all words that are played back by the voice control service to a calling party. In order not to delay the updating a "reserve" voice, however, can be used. Records with reserve voice are automatically registered as residual items/records, to be exchanged for the "ordinary" voice at a coming updating. In one embodiment of the invention the customer can select which voice that shall be used by the application, for instance from a predefined choice. 7. Registration as residual items/records is made automatically, but registration as residual items/records of new words for updating of Telia's OrdBok, is a manual process. In this is included: - transcription recording of sound files handling of matters
- FIG. 2 is shown a brief system description of the automatic updating system for the voice control service DigiTala.
- One embodiment of the system according to the invention includes the following parts:
- HVD Ordinary interception computer
- AUR Automatic updating routine
- the routine for the automatic updating system includes a plurality of steps :
- HVD Loading of an application into the customer's HVD (or the like) that shall transmit the records that shall be included in Telia DigiTala.
- the customer shall in his/her HVD indicate which records that shall be used by Telia DigiTala.
- the application in HVD shall:
- the automatic updating routine is in its turn arranged to, for each customer, automatically:
- the updating system is arranged to for each customer automatically: - derive a customer's HVD-file from the FTP-server, preferably at fixed points of time according to customer profile, or at request; decrypt the file; check format; - convert records to internal format for Telia DigiTala, and give alarm at errors .
- the automatic updating routine is further arranged to;
- the updating system is arranged to, for each record according to customer profiles, automatically: find out whether words in records that shall give transcriptions and sound files exist in Telia's OrdBok; derive records (return values for words, transcriptions and reference to sound file) from Telia's OrdBok for requested words; transmit residual items/records in the HVD-file for handling of residual items/records.
- the automatic updating routine also operates to automatically :
- the updating system is further arranged to:
- AUR therefore works for, for personal records that are lacking transcription or sound file with ordinary voice, to:
- AUR Separate manual systems are needed for handling of residual items/records and handling of matters, as well as the interaction with AUR. This includes customer support, transcription and recording of sound files . From the manual system for handling of residual items/records and handling of matters there are support functions in AUR for handling of the following:
- Catalogue-support supplies to a studio at Prosoft order with manuscript for words that are lacking sound file but have transcription. Manuscript shall be designed for both ordinary voice and reserve voice, because first name and family name shall have the same voice.
- the studio at Prosoft handles recording of sound files according to manuscript for either ordinary voice or reserve voice, according to agreement with Customer-Support.
- the sound files are then transmitted to Catalogue-support in wanted format for verification.
- Catalogue-support at Prosoft is adapted to be used for a great part of the fault management that can occur at the automatic updating.
- Catalogue-support also will handle: - installation of Telia DigiTala (HW and SW) ; installation of order information from customer in the automatic updating. (HVD-provider is responsible for the handling in the customer's interception computer)
- the updating system for a voice control service within telecommunications is characterised in being for the most part automatic.
- the system and the method are utilised to automatically utilise and re-use an existing word database at establishing of a voice service application for a customer.
- the system includes an updating routine, which in said word database searches for the words the customer specifies shall be used in his/her application.
- Existing transcriptions and sound files can at that be utilised to create grammar, i.e. words, or combinations of words, to which the applications shall react, and words that are lacking are automatically registered for handling as residual items/records.
- the updating routine by that produces a system for voice controlled services in which information from a customer is automatically checked to a word database, and where the customer information automatically is supplemented with sound data from the word database to create a customer catalogue for the voice control service. This operates to automatically update words in the grammar that shall be used in a voice-controlled service by sound data that are derived from the word database.
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Telephonic Communication Services (AREA)
- Exchange Systems With Centralized Control (AREA)
Abstract
The invention relates to an updating system for a voice controlled service in a telecommunications system, which service is offered by an operator, and which includes means to recognise and react to a voice command given by a calling party to a customer that subscribes to the service. Customer data with catalogue information are automatically derived from the customer to the operator, and are communicated to the updating system. In the updating system is included a word database with sound data for a voice service, and an updating device arranged to, in the word database, automatically search for and derive words defined in said catalogue information. The updating device creates and updates automatically files by information from the customer's catalogue information with belonging sound data that shall be used for the voice-controlled service.
Description
SYSTEM AND METHOD FOR VOICE-CONTROLLED TELECOMMUNICATIONS SERVICES
Technical field
The present invention relates to a method and device at a telecommunications system that makes possible to automatically update functions on a general level from the user' s interception computer to collection of the information the application need, to be able to handle the user' s catalogue for voice-controlled telecommunications services .
Prior art
Telephone services where a calling party can be forwarded to extensions, give orders for articles/services, provide or obtain information, etc, without the need to talk to a telephone operator are today very frequent. Such services are generally based on that a number of preselections have been programmed by the customer of the service, which pre-selections are presented to the calling party by a recorded voice. These pre-selections then are accessible to the calling party by keying on his/her keyset code telephone. Navigation in a pre-selection programmed exchange service, or, for instance, in a service for ordering of articles, often can be made in several steps with repeated selections by means of keying. Systems that makes possible forwarding to persons only be means of simple keying are normally used by companies, public authorities or organisations, and can wholly or partially replace a manned telephone exchange. However, there is still the weakness that one has to use the keyset to select pre-selection alternative.
In a development of automated telecommunications services, voice recognition devices are used to react to voice commands from the calling party. Such a service further can include the use of sound files, which can be put together to create sentences that are played back as voice answer to the calling party. So called voice- controlled services have several advantages in comparison with key pulsing services. One advantage is that keying is wholly eliminated, with exception of possibly an initial telephone number. Further is, with a voice-controlled service, the possibility obtained to make a multiple of selections by one command, in a way that is not in a simple way possible to achieve with ten buttons of a telephone. It would for instance be possible to, by the command "call Anders Olsson's mobile", provide necessary information about which service one wants to utilise (call), who one wants to call, and where one should try to find him/her (i.e. on his/her mobile telephone). A voice-controlled system consequently can in a better way offer a service that corresponds to a human telephone operator.
Aim of the invention
A lot of voice-controlled telecommunications services today require information from a plurality of different sources. On one hand an operator who offers the voice- controlled service need information about the customer, for instance derived from an external database belonging to the customer. On the other is needed necessary information for speech recognition at the operator's voice service device, i.e. a kind of dictionary. Further are needed, to get a functioning system, information about which words or phrases the system shall react to, and necessary sound files that shall be fed back to the user, i.e. the calling party.
One problem today is that the work to collect this types of information to a common system is not done automatically, but separately from each other and with primarily manual handling.
Another problem with voice-controlled systems is that they require a considerable database with words for recognition, and must be able to understand different pronunciations, to make it possible to utilise the service in an efficient way. The aim of the present invention is provide a voice control system that solves these problems.
Summary of the invention
The invention solves the above mentioned problems by a system and a method to automatically derive necessary information for a voice-controlled service, including sound data. The system includes a general database for words that shall be possible to use in voice-controlled services and where each word is associated with sound data, including return value, transcriptions and sound files. An updating system is arranged to automatically derive information from an external database with words that shall be included in the service, and sound data from the word database. The information from the two databases is distributed automatically to the system for the voice-controlled service. A new database is after that automatically built up that corresponds to the external database, but which now also contains return values and sound files. Separately the application also generates grammar with transmitted return values for words or combinations of words to which the service shall react. In a corresponding way the application generates a dictionary for the different pronunciation of the words in question.
The invention consequently relates to an updating system for a voice-controlled service in a telecommunications system, which service is offered by an operator, for instance Telia, and which includes means to recognise and react to a voice command given by a calling party to a customer who subscribes to the service. The invention includes a customer data device for automatic obtaining of catalogue information from the customer to the operator, which customer data device is connected to the updating system. Further are included in the updating system a word database with sound data for a voice service, and an updating device which is connected to the customer data device and the word database, arranged to, in the word database, automatically search for and derive words defined in said catalogue information. The updating device is further arranged to automatically create and update files with information from the customer' s catalogue information with belonging sound data, which shall be used for the voice-controlled service. The invention is defined in the independent patent claims, whereas advantageous embodiments are presented in the subclaims .
Brief description of the drawings
The invention is described below in detail with reference to the enclosed drawings, of which:
Figure 1 shows a block diagram of the automatic call routine according to the invention, and
Figure 2 shows units and flows in the voice control system according to the invention.
Detailed description of preferred embodiments
A central part of the invention is Telia's OrdBok (Telia's Dictionary) which is arranged to contain sound data for a voice-controlled service. Telia's OrdBok is built up as a general SQL-database for words that shall be possible to use in different voice-controlled services. It contains for each word the following sound data: Return value, transcriptions (descriptions of pronunciation) and sound files. All words have a return value, but there can be more than one word that has the same return value. In that way, for instance the words (names) Eriksson and Erixon, can have the same return value, as a suggestion with the spelling Eriksson. The words have different spelling, but are pronounced in the same way. By return value is meant an index for the word in the database, and can also be called normalised spelling. The updating routine for the voice-controlled service automatically derives information from an external database and sound data from the word database for the words that shall be included in the service.
Information from the two databases is automatically distributed to the system for the voice-controlled service. In ap .' cation for the service is then a new database automat-, -.ally built up that corresponds to the external database but which now also includes return values and sound files. Separately, the application also generates grammar with transmitted return values for words or combination of words, to which the service shall react. In corresponding way the application generates a dictionary for the different pronunciation (transcriptions) of the words in question. Words that cannot be found in Telia's OrdBok are automatically registered as "residual items/records" for manual handling with transcription and recording. When the sound data of the words later have been
entered, the information will be delivered to the service at coming updatings . The need for this handling of residual items/records then will decrease, the larger Telia's OrdBok is becoming. Telia's OrdBok is a word database with necessary sound data for a voice-controlled service. It is worked out to be used together with the automatic updating routine AUR for delivery to Telia's voice control service Telia DigiTala, but due to its construction it is also of current interest for use by other voice controlled services. Telia's OrdBok consequently is database for the words in the application Telia DigiTala that require voice recognition and playback with sound file. At the use of an embodiment of the invention there are three participants in the system: 1) The operator, for instance Telia, which provides the voice control service, 2) The customer, which subscribes to the voice control service, for instance a company that utilises the service as a telephone exchange, and 3) a user or a calling party who, by means of voice- commands, can utilise the voice control service for wanted purpose.
A customer defines necessary data in a file on an interception computer HVD . This interception computer is normally located at the customer's, and is arranged to tra "^ z HVD-files to the operator. This is done, prefer-..- iy, over Internet to an FTP-server at the operator's, which server is included in the voice control system. To make it possible to establish and use the voice control service is required, in a preferred embodiment, that the following words are included in the word database from the customers HVD-files:
First name Family name - Department, division, or place
Function (such as reception, porter's office etc) .
In the database is also included a master catalogue for sound files, preferably files of PC-format. In Telia's OrdBok can for each word be found:
- word (according to record, spelling according to HVD) return value (normalised spelling) 1)
Both these are preferably lower-case letters with underscore instead of space type of word (for instance first name, family name, department etc.) transcription of normal pronunciation transcription of alternative pronunciation (one or more) reference to sound file for normal pronunciation. Also references to sound files for alternative pronunciation can be included. voice (ordinary or reserve voice) status. In status is included date for AUR's obtaining of sound file to respective platform (PC or SRP) .
Telia's OrdBok (dictionary) is arranged to:
store records with word, transcription (with alternative pronunciation) and references to sound file; pr 'ide information to AUR about which words in record that lacks transcription and sound file; transmit to AUR requested information (return value for words, transcriptions and reference to sound file)
In Telia's OrdBok the transcriptions are unique within each type of word for each return value, i.e. a transcription exists only one time in Telia's OrdBok for a type of word. The reason for the addition "within each type of word" depends on that a word can exist as different types of word, for instance Helen as first name and family
name. Telia's OrdBok contains words with both transcription and sound file, and it is updated running. New words are transcribed and a studio attends to that new-recorded sound files are available in the updating system.
The words that are entered in Telia's OrdBok for Telia DigiTala correspond to the text that is in respective field in the customer's HVD-file as above. The words that are entered correspond to the text the customer has written in corresponding fields. This means that double first names are both handled as one word, i.e. Sven Ingvar and Sven- Ingvar are entered in Telia's OrdBok as one word. In corresponding way double family names are handled, Petterson Lundstrδm and Petterson-Lundstrδm are entered as one word. Alternative spellings shall in Telia's OrdBok be handled so that each spelling is a new word. In Telia's OrdBok then the same normal spelling (return value) , transcriptions and reference to sound file shall be used for words with alternative spellings. In AUR alternative spellings are entered at the same time as a new word is to be entered in Telia's OrdBok, by search in a specific file with alternative spellings of first names and family names.
Telia DigiTala is a voice-controlled service which makes it possible for a user to automatically be connected to an e tension in an exchange by using the voice only. Catalogue information for the customer is derived from the customer' s interception computer and is updated for the service Telia DigiTala by a separate updating system according to the invention. The system consists primarily of an automatic updating system AUR (Automatic Updating Routine) .
AUR is arranged to reuse recorded sound files and transcriptions collected in a separate database, preferably Telia's OrdBok. As complement there is a system for manual
updating of the residual items/records that are lacking sound data in Telia's OrdBok. The updating system provides both information for applications on PC-platform and SRP- platform. According to the invention the updating system AUR for Telia DigiTala attends to that the customer's catalogue in Telia DigiTala is automatically updated according to the customer's wishes. In a preferred embodiment this is done at fixed points of time. Alternatively the updating can be made at request from the customer, or the operator. The basis of the updating is the customer's HVD and the customer's request about which information in HVD that shall be accessible in Telia DigiTala. The updated information is made accessible to the application on the platform, PC or SRP. Separately, outside the updating system, the information is distributed within the platforms.
In Figure 1 is shown a block diagram over the updating process of AUR, for which the following principles apply:
1. Customer data registered as records in an HVD-file are transmitted to the updating system, preferably from the customer' s interception computer to an FTP-server over Internet, as has previously been described, from where the updatir.σ system derives the HVD-file.
2. In the cases words in records in HVD exist in Telia's OrdBok, the files that the application Telia DigiTala requires for updating of a compiler are created automatically.
3. For the records in HVD that contains words that do not exist in Telia's OrdBok, the record is left automatically for handling of residual items/records with transcription and recording. A registration for residual items/records
is at that made for a correction measure with manual handling. At a later collection of HVD-files, the previously lacked records will be in the updating.
4. If first name and family name are lacking in Telia's OrdBok, lacking words are automatically registered as residual items/records. The personal record then will not be included in TeliaDigitala .
5. If the word for a group membership does not exist in Telia's Ordbok, lacking word is automatically registered as residual item/record, but the personal record will yet be included in Telia DigiTala. In a preferred embodiment it then will be possible to call such a record, but if there are a plurality of persons with the same pronunciation of name, the application is arranged to connect to telephone operator until the word of the group membership has been entered in Telia's Ordbok. For all records there consequently shall be a group membership to make it possible to distinguish between persons with the same pronunciation of name. It can, for instance, be division, department or place, which is determined in connection with the ordering of the service.
6. The new updating for new records is primarily deliver d by an "ordinary" voice, i.e. a voice which is used for all words that are played back by the voice control service to a calling party. In order not to delay the updating a "reserve" voice, however, can be used. Records with reserve voice are automatically registered as residual items/records, to be exchanged for the "ordinary" voice at a coming updating. In one embodiment of the invention the customer can select which voice that shall be used by the application, for instance from a predefined choice.
7. Registration as residual items/records is made automatically, but registration as residual items/records of new words for updating of Telia's OrdBok, is a manual process. In this is included: - transcription recording of sound files handling of matters
In Figure 2 is shown a brief system description of the automatic updating system for the voice control service DigiTala. One embodiment of the system according to the invention includes the following parts:
1. Ordinary interception computer (HVD) supplemented with application for selection, encryption and transmission.
2. FTP-server.
3. Automatic updating routine (AUR) to generate data for Telia DigiTala and for the handling of the residual items/records.
4. Telia OrdBok with transcription and references to sound files for words. 5. Platform with customer system for Telia DigiTala with distribution application for data from AUR. 6. Manual handling for catalogue support, as "back of.-ce" .
7. Manual system for transcription.
8. Manuel system for handling of recordings.
According to the present invention, the routine for the automatic updating system includes a plurality of steps :
Loading of an application into the customer's HVD (or the like) that shall transmit the records that shall be included in Telia DigiTala. The customer shall in his/her
HVD indicate which records that shall be used by Telia DigiTala. The application in HVD shall:
attend to select records in HVD that the customer has indicated shall be included in Telia DigiTala; create an HVD-file with data for Telia DigiTala; encrypt the file; transmit the file to an FTP-server at the operator, or the service provider, preferably at fixed points of time. The FTP-server is used for buffering of customers transmitted HVD-files on account for respective customer.
The automatic updating routine is in its turn arranged to, for each customer, automatically:
- derive data from the FTP-server. At that, the updating system is arranged to for each customer automatically: - derive a customer's HVD-file from the FTP-server, preferably at fixed points of time according to customer profile, or at request; decrypt the file; check format; - convert records to internal format for Telia DigiTala, and give alarm at errors .
The automatic updating routine is further arranged to;
handle databases for sound files and transcription. At that, the updating system is arranged to, for each record according to customer profiles, automatically: find out whether words in records that shall give transcriptions and sound files exist in Telia's OrdBok;
derive records (return values for words, transcriptions and reference to sound file) from Telia's OrdBok for requested words; transmit residual items/records in the HVD-file for handling of residual items/records.
The automatic updating routine also operates to automatically :
- create files for delivery to application. This includes to create necessary files to update the application Telia DigiTala (phonebook, grammar, attendant etc) - pack the files for each customer, compress and make them accessible to be derived by the distribution application for existing platform. Basis for creation of files is customer profile and the information in HVD-file. Transcriptions for words and references to sound files are derived from Telia's OrdBok. New sound files are packed automatically and are made accessible to be derived by the distribution system to the platform. For sound files to the SRP-platform, AUR converts automatically these to VSN-format before packing.
The updating system is further arranged to:
- transmit records or words in records that are lacking in the word database for manual handling of residual items/records. AUR therefore works for, for personal records that are lacking transcription or sound file with ordinary voice, to:
store the records of all customers from the HVD-
files of the customers that are lacking transcription and sound files; give the manual handling system, preferably managed by Telia Prosoft, information about what shall be transcribed and recorded; handle the updating of Telia's OrdBok. Complete personal record, but with sound file with reserve voice, is handled by AUR for new recording of ordinary voice .
Separate manual systems are needed for handling of residual items/records and handling of matters, as well as the interaction with AUR. This includes customer support, transcription and recording of sound files . From the manual system for handling of residual items/records and handling of matters there are support functions in AUR for handling of the following:
Editing of customer profile; - Fetching of basic data for word that shall be transcribed;
Production of manuscript for recording;
Entering of verified sound files in AUR;
Population of Telia's OrdBok according to the handling of residual items/records.
Replacement of faulty sound file;
Population of Telia's OrdBok with new words separate from the handling of residual items/records;
Manual operation of AUR;
The following control functions also exist in AUR:
Control of records in the customer's HVD with persons who have names the pronunciation of which are identical;
Control of the number of records in the customer's HVD in comparison with customer profile;
For customer support there is in the manual system Catalogue-support, the support unit with back-office function that shall attend to operation of the customer' s catalogue in Telia DigiTala for the handling of the residual items/records, the handling of matters, and the handling of errors . In AUR there are functions to support Catalogue-support. This includes:
production of basic data for transcription and manuscript for recording from the handling of the residual items/records of AUR; control listening to verify new recorded sound files; population of Telia's OrdBok according to made transcriptions and recordings.
In AUR there are functions for supporting the manual transcription and recording of new words. Catalogue-support supplies to a studio at Prosoft order with manuscript for words that are lacking sound file but have transcription. Manuscript shall be designed for both ordinary voice and reserve voice, because first name and family name shall have the same voice. The studio at Prosoft handles recording of sound files according to manuscript for either ordinary voice or reserve voice, according to agreement with Customer-Support. The sound files are then transmitted to Catalogue-support in wanted format for verification. Catalogue-support at Prosoft is adapted to be used for a great part of the fault management that can occur at the automatic updating. Catalogue-support also will handle: - installation of Telia DigiTala (HW and SW) ; installation of order information from customer in the automatic updating. (HVD-provider is responsible for the handling in the customer's interception computer)
The updating system for a voice control service within telecommunications according to the present invention is
characterised in being for the most part automatic. The system and the method are utilised to automatically utilise and re-use an existing word database at establishing of a voice service application for a customer. The system includes an updating routine, which in said word database searches for the words the customer specifies shall be used in his/her application. Existing transcriptions and sound files can at that be utilised to create grammar, i.e. words, or combinations of words, to which the applications shall react, and words that are lacking are automatically registered for handling as residual items/records. The updating routine by that produces a system for voice controlled services in which information from a customer is automatically checked to a word database, and where the customer information automatically is supplemented with sound data from the word database to create a customer catalogue for the voice control service. This operates to automatically update words in the grammar that shall be used in a voice-controlled service by sound data that are derived from the word database.
Above preferred embodiments of the invention have been described, whereas the scope of protection is defined by enclosed patent claims.
Claims
1. An updating system for a voice controlled service in a telecommunications system, which service is offered by an operator and includes means to recognise and react to a voice command given by a calling party to a customer who subscribes to the service, c h a r a c t e r i s e d in that a customer data device, for automatic collection of catalogue information from the customer to the operator, is connected to the updating system, and that the updating system includes a word database with sound data for a voice service, and an updating device connected to the customer data device and the word database, arranged to, in the word database, automatically search for and derive words defined in said catalogue information.
2. The updating system as claimed in patent claim 1, where the updating device further includes a compiler arranged to automatically create and update files for the voice control service by information from the customer' s catalogue information with belonging sound data.
3. The updating system as claimed in patent claim 2, where the updating device is connected to a voice service database for the voice control service, which is arranged to be automatically updated by files from the updating device .
4. The updating system as claimed in any of the previous patent claims, where the updating device further is arranged to register as residual items/records words or records in the customer's catalogue information that are lacking in the word database.
5. The updating system as claimed in patent claim 4, further including a manual handling system connected to the updating device, which handling system is arranged to handle as residual items/records registered words, and update the word database.
6. The updating system as claimed in any of the previous patent claims, where sound data in said word database for each word includes return value, transcriptions and sound files .
7. The updating system as claimed in any of the previous patent claims, where said customer data device includes an interception computer at the customer and an FTP-server at the operator, to which catalogue information is automatically transmitted from the interception computer over Internet.
8. The updating system as claimed in patent claim 7, where said interception computer is supplemented with application for selection of information that shall be included in the catalogue information, encryption and transmission of files to the FTP-server via Internet.
9. The updating system as claimed in patent claim 8, where said interception computer is arranged to transmit a file with catalogue information to the FTP-server at fixed points of time.
10. The updating system as claimed in patent claim 8, where said interception computer is arranged to transmit a file with catalogue information to the FTP-server at request .
11. A method for an updating system for a voice-controlled service in a telecommunications system, which service is offered by an operator to recognise and react to a voice command given by a calling party to a customer who subscribes to the service, c h a r a c t e r i s e d in the steps to automatically derive catalogue information from the customer to the operator with a customer data device connected to the updating system; search for and derive words defined in said catalogue information in a word database including sound data for a voice service, by an updating device in the updating system.
12. The method as claimed in patent claim 11, further including the step to, in a compiler in the updating device, automatically create and update files for the voice service by information from the customer' s catalogue information supplemented with belonging sound data.
13. The method as claimed in patent claim 12, further including the step to automatically transmit said files to a voice service database for the voice service.
14. The method as claimed in any of the patent claims Ills, further including the step to, in the updating device, automatically register as residual items/records words or records in the customer' s catalogue information that are lacking in the word database.
15. The method as claimed in patent claim 14, further including the step to automatically transmit words registered as residual items/records to a manual handling system connected to the updating device, to handle words registered as residual items/records in the handling system, and to update the word database.
16. The method as claimed in any of the patent claims 11- 15, where sound data in said word database for each word includes return value, transcriptions and sound files.
17. The method as claimed in any of the patent claims 11- 16, where the collection step includes the steps to, at the customer side, select information that shall be included in the catalogue information; create and encrypt a file with this catalogue information; transmit the file to the updating system at the operator via Internet, and to, at the operator's side, buffer the customer's file on an FTP-server; automatically derive a customer' s file from the FTP- server; decrypt the file; - control the format of the file; convert records in the catalogue information of the file to internal format for the voice service; give alarm at fault/error.
18. The method as claimed in patent claim 17, where transmission of a file with catalogue information to the FTP-server is made automatically at fixed points of time.
19. The method as claimed in patent claim 17, where transmission of a file with catalogue information to the FTP-server is made at request.
20. A method for an updating system for a voice controlled service in a telecommunications system, which service is offered by an operator to recognise and react to a voice command given by a calling party to a customer who subscribes to the service, c h a r a c t e r i s e d in the steps to load an application into a customer's interception computer (HVD) that shall transmit the records that shall be included in the voice service, which application serves to select records that shall be used by the voice service, and which have been indicated by the customer; create an HVD-file by data for the voice service; transmit the file to an FTP-server at the operator, or the service provider, at which an automatic updating routine in the updating system is arranged to - derive a customer's HVD-file from the FTP-server at fixed points of time according to customer profile, or at request; find out whether words in records in the HVD-file, that shall give transcriptions and sound files, are in a word database in the updating system; derive records with return value for words, transcriptions and reference to sound files for requested words in the word database; transmit residual items/records in the HVD-file for handling of residual items/records; create necessary files to update an application by a voice controlled service; packet these files for each customer, compress and make them accessible for collection by a distribution application for existing platform (PC, SRP) .
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| SE9902231A SE521172C2 (en) | 1999-06-14 | 1999-06-14 | Systems and method for voice-controlled telecommunications services |
| SE9902231-1 | 1999-06-14 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2000078024A1 true WO2000078024A1 (en) | 2000-12-21 |
Family
ID=20416059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/SE2000/001240 Ceased WO2000078024A1 (en) | 1999-06-14 | 2000-06-13 | System and method for voice-controlled telecommunications services |
Country Status (2)
| Country | Link |
|---|---|
| SE (1) | SE521172C2 (en) |
| WO (1) | WO2000078024A1 (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1997019545A1 (en) * | 1995-11-17 | 1997-05-29 | At & T Corp. | Automatic vocabulary generation for telecommunications network-based voice-dialing |
| WO1998013992A1 (en) * | 1996-09-25 | 1998-04-02 | At & T Corp. | Voice dialing system |
-
1999
- 1999-06-14 SE SE9902231A patent/SE521172C2/en not_active IP Right Cessation
-
2000
- 2000-06-13 WO PCT/SE2000/001240 patent/WO2000078024A1/en not_active Ceased
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1997019545A1 (en) * | 1995-11-17 | 1997-05-29 | At & T Corp. | Automatic vocabulary generation for telecommunications network-based voice-dialing |
| WO1998013992A1 (en) * | 1996-09-25 | 1998-04-02 | At & T Corp. | Voice dialing system |
Also Published As
| Publication number | Publication date |
|---|---|
| SE9902231L (en) | 2000-12-15 |
| SE521172C2 (en) | 2003-10-07 |
| SE9902231D0 (en) | 1999-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US5940476A (en) | System and method for identifying an unidentified caller | |
| US6941273B1 (en) | Telephony-data application interface apparatus and method for multi-modal access to data applications | |
| EP0735736B1 (en) | Method for automatic speech recognition of arbitrary spoken words | |
| US6018568A (en) | Voice dialing system | |
| US5493606A (en) | Multi-lingual prompt management system for a network applications platform | |
| US6885733B2 (en) | Method of providing a user interface for audio telecommunications systems | |
| US5485507A (en) | Integrated commissary system | |
| US6937986B2 (en) | Automatic dynamic speech recognition vocabulary based on external sources of information | |
| US7283973B1 (en) | Multi-modal voice-enabled content access and delivery system | |
| US7324942B1 (en) | System and method for interactive voice services using markup language with N-best filter element | |
| US6195417B1 (en) | Automated system for accessing speech-based information | |
| US20080187109A1 (en) | Audio archive generation and presentation | |
| WO2001043024A1 (en) | Computer accounting method using natural language speech recognition | |
| WO2008137327A1 (en) | Automated attendant grammar tuning | |
| US20020169615A1 (en) | Computerized voice-controlled system for compiling quality control data | |
| EP0872102A4 (en) | Interactive medication data telephony system | |
| US20240176957A1 (en) | Systems and methods for inserting dialogue into a query response | |
| US20050114139A1 (en) | Method of operating a speech dialog system | |
| US20040193603A1 (en) | Technique for effectively searching for information in response to requests in information assistance service | |
| US7689425B2 (en) | Quality of service call routing system using counselor and speech recognition engine and method thereof | |
| CA2500267C (en) | Prepaid phone card service using speech recognition | |
| WO2000078024A1 (en) | System and method for voice-controlled telecommunications services | |
| US20060093103A1 (en) | Technique for generating and accessing organized information through an information assistance service | |
| US6502075B1 (en) | Auto attendant having natural names database library | |
| US8626523B1 (en) | Patient voice check-in system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A1 Designated state(s): EE LT LV NO US |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
| 122 | Ep: pct application non-entry in european phase |