Applications Server and Method

• US 20080262848A1
• Filed: 01/03/2006
• Published: 10/23/2008
• Est. Priority Date: 01/06/2005
• Status: Abandoned Application

• Current Assignee: Orange S.A.
• Original Assignee: Orange S.A.
• Inventors: Adamsky, Robert, Howe, Gregory, Shienbrood, Eric, Pelland, David
• Application Number:
  US11/794,877
• Publication Number:
  US 20080262848A1
• Time in Patent Office: Days
• US Class Current: 704/275
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 2203/355 Interactive dialogue design tools, features or methods
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00

Abstract

A speech applications server is arranged to provide a user driven service in accordance with an application program in response to user commands for selecting service options. The user is prompted by audio prompts to issue the user commands. The application program comprises a state machine operable to determine a state of the application program from one of a predetermined set of states defining a logical procedure through the user selected service options, transitions between states being determined in accordance with logical conditions to be satisfied in order to change between one state of the set and another state of the set. The logical conditions include whether a user has provided one of a set of possible commands. A prompt selection engine is operable to generate the audio prompts for prompting the commands from the user in accordance with predetermined rules. The prompt selected by the prompt selection engine is determined at run-time. Since the state machine and the prompt selection engine are separate entities and the prompts to be selected are determined at run-time, it is possible to effect a change to the prompt selection engine without influencing the operation of the state machine, enabling different customisations to be provided for the same user driven services, in particular this allows multilingual support, with the possibility of providing rules to adapt the prompt structure allowing for grammatical differences between to languages to be taken into account thus providing higher quality multiple language support.

Claims

1. A speech applications server operable to provide a user driven service in response to user commands for selecting service options, the user commands being prompted by audio prompts, the speech applications server comprising

at least one state machine operable by a state machine engine to determine a state of an application program from one of a predetermined set of states defining a logical procedure through the user selected service options, transitions between states being determined in accordance with logical conditions to be satisfied in order to change between one state of the set and another state of the set, the logical conditions including whether a user has provided one of a set of possible commands, and

a prompt selection rule set operable by a prompt selection engine to generate the audio prompts for prompting the commands from the user in accordance with predetermined rules, wherein the prompt selected by the prompt selection engine is determined at run-time and the at least one state machine of the application program is defined separately from the prompt selection rule set to the effect that a change can be made to the prompt selection rule set defining a dialogue generated by the prompt selection engine for the user driven service independently from the operation of the state machine,

wherein the state machine is defined using a mark-up language, the mark-up language including

a form instruction for defining a set of the possible states of the application program within a dialog form, each form state including at least one situation identifier for identifying at least one current state and a logical condition for changing to the following state, and a request to the prompt selection engine to generate a request for a user command for satisfying the logical condition for the application program to change to the following state, and

wherein the prompt selection rule set is defined using a mark-up language, the mark-up language defining for each situation identifier the set of possible prompts which may be provided to a user, and wherein,

the application program is formed by translating the mark-up language for the state machine and the mark-up language for the prompt selection rule sets into executable code, the executable code being operable to generate the VoiceXML mark-up language, which when communicated to a telephony platform provides the user driven service.

2. The speech applications server as claimed in claim 1, comprising a command recognition engine, the command recognition engine including a speech recogniser, the speech recogniser being operable to provide the command recognition engine with the set of possible user commands which may be received from the user with respect to the logical condition for changing from a current one of the predetermined set of states to another of the states of the state machine, the command recognition engine being operable

to analyse commands provided by the user with respect to the set of possible commands provided by the speech recogniser, and

to provide the state machine with an estimate of one of the set of possible commands which the user provided, the state machine being operable to change state in response to the estimated user command.

3. The speech applications server as claimed in claim 2, wherein the command recognition engine is operable to determine a confidence level for the estimated user command, the state machine identifying the change of state from the estimated user command in combination with the determined confidence level.

4. The speech applications server as claimed in claim 3, wherein the user commands include voice commands, the speech recognition processor being operable to generate the confidence levels for the estimate of the possible voice commands.

5. The speech applications server as claimed in claim 1, wherein the application program is operable to generate a mark-up language page in accordance with a current state of the application program as determined by the state machine, the mark-up language page including universal resource locations (URLs) defining a location for data files providing the audio prompts, the speech grammars, and the DTMF grammars.

6. The speech applications server as claimed in claim 5 wherein the mark-up language is VoiceXML.

7. The speech applications server as claimed in claim 2, wherein command recognition grammars are specified using the mark-up language.

8. The speech applications server as claimed in claim 5, comprising a web server operable to receive the mark-up language page and to deploy separately the mark-up language page to a telephony platform.

26. The speech applications server as claimed in claim 2, wherein the application program is operable to generate a mark-up language page in accordance with a current state of the application program as determined by the state machine, the mark-up language page including universal resource locations (URLs) defining a location for data files providing the audio prompts, speech grammars, and DTMF grammars.

27. The speech applications server as claimed in claim 3, wherein the application program is operable to generate a mark-up language page in accordance with a current state of the application program as determined by the state machine, the mark-up language page including universal resource locations (URLs) defining a location for data files providing the audio prompts, speech grammars, and DTMF grammars.

28. The speech applications server as claimed in claim 4, wherein the application program is operable to generate a mark-up language page in accordance with a current state of the application program as determined by the state machine, the mark-up language page including universal resource locations (URLs) defining a location for data files providing the audio prompts, speech grammars, and DTMF grammars.

29. The speech applications server as claimed in claim 3, wherein command recognition grammars are specified using the mark-up language.

30. The speech applications server as claimed in claim 4, wherein command recognition grammars are specified using the mark-up language.

31. The speech applications server as claimed in claim 5, wherein command recognition grammars are specified using the mark-up language.

32. The speech applications server as claimed in claim 6, wherein command recognition grammars are specified using the mark-up language.

33. The speech applications server as claimed in claim 6, comprising a web server operable to receive the mark-up language page and to deploy separately the mark-up language page to a telephony platform.

34. The speech applications server as claimed in claim 7, comprising a web server operable to receive the mark-up language page and to deploy separately the mark-up language page to a telephony platform.

9. An application program operable to provide a user driven service in response to user commands for selecting service options, the user commands being prompted by audio prompts, the application program comprising

at least one state machine operable by a state machine engine to determine a state of the application program from one of a predetermined set of states defining a logical procedure through the user selected service options, transitions between states being determined in accordance with logical conditions to be satisfied in order to change between one state of the set and another state of the set, the logical conditions including whether a user has provided one of a set of possible commands, and

a prompt selection rule set operable by a prompt selection engine to generate the audio prompts for prompting the commands from the user in accordance with predetermined rules, wherein the prompt selected by the prompt selection engine is determined at run-time and the at least one state machine of the application program is defined separately from the prompt selection rule set to the effect that a change can be made to the prompt selection rule set defining a dialogue generated by the prompt selection engine for the user driven service independently from the operation of the state machine,

wherein the state machine is defined using a mark-up language, the mark-up language including

a form instruction for defining a set of the possible states of the application program within a dialog form, each form state including at least one situation identifier for identifying at least one current state and a logical condition for changing to the following state, and a request to the prompt selection engine to generate a request for a user command for satisfying the logical condition for the application program to change to the following state, and

wherein the prompt selection rule set is defined using a mark-up language, the mark-up language defining for each situation identifier the set of possible prompts which may be provided to a user, and wherein,

the application program is formed by translating the mark-up language for the state machine and the mark-up language for the prompt selection rule sets into executable code, the executable code being operable to generate the VoiceXML mark-up language, which when communicated to a telephony platform provides the user driven service.

10. The application program as claimed in claim 9, wherein the state machine is responsive to an estimate of one of the possible commands which the user may have provided in accordance with the logical condition for changing from a current one of the predetermined set of states to another of the states of the application program to change to the other state, the estimate of the user command being provided by a command recognition engine, the command recognition engine including a speech recogniser, the speech recogniser being operable to provide the command recognition engine with the set of possible user commands which may be received from the user with respect to the logical conditions for changing from the current state to the other state, the command recognition engine being operable to analyse the user commands with respect to the possible set of user commands which the user may have provided, wherein the state machine is operable to change state in response to the estimated user command in combination with the logical condition associated with the change of state.

11. The application program as claimed in claim 10, wherein the state machine is operable to identify the change of state from the estimated user command in combination with a determined confidence level, the determined confidence level being provided by the command recognition engine.

12. The application program as claimed in claim 11, wherein the user commands include voice commands, the speech recognition processor being operable to generate the confidence levels for the estimate of the possible voice commands.

13. A system for providing a user driven service, the system comprising a speech applications server, a telephony platform and a user equipment,

the speech applications server being operable to provide a user driven service in response to user commands for selecting service options, the user commands being prompted by audio prompts, the speech applications server comprising

at least one state machine operable by a state machine engine to determine a state of an application program from one of a predetermined set of states defining a logical procedure though the user selected service options, transitions between states being determined in accordance with logical conditions to be satisfied in order to change between one state of the set and another state of the set, the logical conditions including whether a user has provided one of a set of possible commands, and

a prompt selection rule set operable by a prompt selection engine to generate the audio prompts for prompting the commands from the user in accordance with predetermined rules, the telephony platform being operable

to receive data representing the audio prompts from the applications server, and

to communicate the audio prompt data to the user equipment, and

to receive data representative of the possible commands from the user equipment, and

to communicate the possible commands to the applications server, wherein the prompt selected by the prompt selection engine is determined at run-time and the at least one state machine of the application program is defined separately from the prompt selection rule set to the effect that a change can be made to the prompt selection rule set defining a dialogue generated by the prompt selection engine for the user driven service independently from the operation of the state machine,

wherein the state machine is defined using a mark-up language, the mark-up language including

a form instruction for defining a set of the possible states of the application program within a dialog form, each form state including at least one situation identifier for identifying at least one current state and a logical condition for changing to the following state, and a request to the prompt selection engine to generate a request for a user command for satisfying the logical condition for the application program to change to the following state, and

wherein the prompt selection rule set is defined using a mark-up language, the mark-up language defining for each situation identifier the set of possible prompts which may be provided to a user, and wherein,

the application program is formed by translating the mark-up language for the state machine and the mark-up language for the prompt selection rule sets into executable code, the executable code being operable to generate the VoiceXML mark-up language, which when communicated to a telephony platform provides the user driven service.

14. A method for providing a user driven service, the service being provided in response to user commands for selecting service options, the user commands being prompted by audio prompts, the method comprising

determining a state of the application program from one of a predetermined set of states defining a logical procedure through the user selected service options, transitions between states being determined in accordance with logical conditions to be satisfied in order to change between one state of the set and another state of the set, the logical conditions including whether a user has provided one of a set of possible commands, and

selecting the audio prompts for prompting the possible commands from the user in accordance with a predetermined rule set, wherein the prompts are generated at run-time and the states of the application program are defined separately from the predetermined prompt selection rule set to the effect that a change can be made to the prompt selection rule set defining a dialogue for the user driven service independently from determining the state of the application program,

wherein the state machine is defined using a mark-up language, the mark-up language including

a form instruction for defining a set of the possible states of the application program within a dialog form, each form state including at least one situation identifier for identifying at least one current state and a logical condition for changing to the following state, and a request to the prompt selection engine to generate a request for a user command for satisfying the logical condition for the application program to change to the following state, and

wherein the prompt selection rule set is defined using a mark-up language, the mark-up language defining for each situation identifier the set of possible prompts which may be provided to a user, and wherein,

the application program is formed by translating the mark-up language for the state machine and the mark-up language for the prompt selection rule sets into executable code, the executable code being operable to generate the VoiceXML mark-up language, which when communicated to a telephony platform provides the user driven service.

15. The method as claimed in claim 14, comprising

analysing the user commands with respect to the possible set of user commands which the user may have provided, the possible set of the user commands being determined in accordance with the logical condition for changing from a current one of the predetermined set of states to another of the states to which the state machine may change,

estimating of one of the possible commands which the user provided, and

changing state in response to the estimated user command in combination with the logical condition associated with the change of state.

16. The method as claimed in claim 15, comprising

determining a confidence level for the estimated user command, the state machine identifying the change of state from the estimated user command in combination with the determined confidence level.

17. The method as claimed in claim 16, wherein the user commands include voice commands, the generating the confidence levels for the estimate of the voice command, the possible user commands being possible voice commands.

18. An extended mark-up language for defining a user driven service, the user driven service being provided in response to user commands for selecting service options, the user commands being prompted by audio prompts, the mark-up language comprising

a form instruction for defining a set of possible states of the service in a dialog form, each form including at least one situation identifier for identifying at least one current state and the logical condition for changing to that state, and a prompt request to generate a request for a user command for satisfying the logical condition for changing to a following state.

19. The extended mark-up language as claimed in claim 18, wherein each of the identified situations within a form is identified by a possible combination of slots corresponding to one of the set of possible commands associated with the situation, and the prompt request is arranged to prompt for the user commands to fill the slots for a set of possible following states.

20. The extended mark-up language as claimed in claim 18 wherein each of the identified situations within a form is further identified by a Boolean expression corresponding to logical conditions necessary for the situation to be selected as the current one.

21. The extended mark-up language as claimed in claim 18, wherein each identified situation includes a suggested prompt to be played, an action to be performed, and instructions for setting up the transition to a possible new set of states.

22. The extended mark-up language as claimed in claim 18, wherein the prompt request is associated with one or more rules, each rule consisting of the situation identifier, a predicate defining preconditions for executing the prompt request and a prompt sequence.

35. The extended mark-up language as claimed in claim 19, wherein each of the identified situations within a form is further identified by a Boolean expression corresponding to logical conditions necessary for the situation to be selected as the current one.

36. The extended mark-up language as claimed in claim 19, wherein each identified situation includes a suggested prompt to be played, an action to be performed, and instructions for setting up the transition to a possible new set of states.

37. The extended mark-up language as claimed in claim 20, wherein each identified situation includes a suggested prompt to be played, an action to be performed, and instructions for setting up the transition to a possible new set of states.

38. The extended mark-up language as claimed in claim 19, wherein the prompt request is associated with one or more rules, each rule consisting of the situation identifier, a predicate defining preconditions for executing the prompt request and a prompt sequence.

39. The extended mark-up language as claimed in claim 20, wherein the prompt request is associated with one or more rules, each rule consisting of the situation identifier, a predicate defining preconditions for executing the prompt request and a prompt sequence.

40. The extended mark-up language as claimed in claim 21, wherein the prompt request is associated with one or more rules, each rule consisting of the situation identifier, a predicate defining preconditions for executing the prompt request and a prompt sequence.

23. A computer program which when loaded on to a data processor causes the data processor to perform a method for providing a user driven service, the service being provided in response to user commands for selecting service options, the user commands being prompted by audio prompts, the method comprising

determining a state of the application program from one of a predetermined set of states defining a logical procedure through the user selected service options, transitions between states being determined in accordance with logical conditions to be satisfied in order to change between one state of the set and another state of the set, the logical conditions including whether a user has provided one of a set of possible commands, and

generating the audio prompts for prompting the possible commands from the user in accordance with predetermined rules, wherein the prompts are generated at run-time and the states of the application program are defined separately from the predetermined prompt selection rule set to the effect that a change can be made to the prompt selection rule set defining a dialogue for the user driven service independently from determining the state of the application program,

wherein the state machine is defined using a mark-up language, the mark-up language including

a form instruction for defining a set of the possible states of the application program within a dialog form, each form state including at least one situation identifier for identifying at least one current state and a logical condition for changing to the following state, and a request to the prompt selection engine to generate a request for a user command for satisfying the logical condition for the application program to change to the following state, and

wherein the prompt selection rule set is defined using a mark-up language, the mark-up language defining for each situation identifier the set of possible prompts which may be provided to a user, and wherein,

the application program is formed by translating the mark-up language for the state machine and the mark-up language for the prompt selection rule sets into executable code, the executable code being operable to generate the VoiceXML mark-up language, which when communicated to a telephony platform provides the user driven service.

24. (canceled)

25. An apparatus for providing a user driven service, the service being provided in response to user commands for selecting service options, the user commands being prompted by audio prompts, the apparatus comprising

means for determining a state of the application program from one of a predetermined set of states defining a logical procedure through the user selected service options, transitions between states being determined in accordance with logical conditions to be satisfied in order to change between one state of the set and another state of the set, the logical conditions including whether a user has provided one of a set of possible commands, and

means for selecting the audio prompts for prompting the possible commands from the user in accordance with a predetermined rule set, wherein the prompts are generated at run-time and the states of the application program are defined separately from the predetermined prompt selection rule set to the effect that a change can be made to the prompt selection rule set defining a dialogue for the user driven service independently from determining the state of the application program,

wherein the state machine is defined using a mark-up language, the mark-up language including

a form instruction for defining a set of the possible states of the application program within a dialog form, each form state including at least one situation identifier for identifying at least one current state and a logical condition for changing to the following state, and a request to the prompt selection engine to generate a request for a user command for satisfying the logical condition for the application program to change to the following state, and

wherein the prompt selection rule set is defined using a mark-up language, the mark-up language defining for each situation identifier the set of possible prompts which may be provided to a user, and wherein,

the application program is formed by translating the mark-up language for the state machine and the mark-up language for the prompt selection rule sets into executable code, the executable code being operable to generate the VoiceXML mark-up language, which when communicated to a telephony platform provides the user driven service.