METHOD AND SYSTEM FOR PROCESSING MULTIPLE DIALOG SESSIONS IN PARALLEL

• US 20090125813A1
• Filed: 11/07/2008
• Published: 05/14/2009
• Est. Priority Date: 11/09/2007
• Status: Abandoned Application

• Current Assignee: Robert Bosch GmbH
• Original Assignee: Robert Bosch GmbH
• Inventors: Shen, Zhongnan, Weng, Fuliang, Meng, Yao
• Application Number:
  US12/266,827
• Publication Number:
  US 20090125813A1
• Time in Patent Office: Days
• US Class Current: 715/728
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/285 Memory allocation or algorithm optimisation to reduce hardware requirements
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

A dialog system and method may generate and maintain in parallel multiple dialog sessions, determine to which dialog session a user speech input applies, selectively provide control to one of the dialog sessions, at any one time, to output data to the user, synchronize multiple dialog sessions, and support user interruptions at any time during the dialog sessions.

Claims

1. A method, comprising:

receiving spoken dialog from a user at a user interface;

generating a plurality of user-initiated actions by a processor based on the received spoken dialog; and

distributing the plurality of user-initiated actions by the processor to a plurality of plan sessions configured to simultaneously execute under control of the processor, each plan session configured to;

store in a memory a queue of distributed user-initiated actions, and execute plan scripts to perform the distributed user-initiated actions.

2. The method of claim 1, wherein a first user-initiated action is distributed to a first plan session, a second user-initiated action is distributed to a second plan session, and the first and second plan sessions simultaneously execute to perform the first and second user-initiated actions.

3. The method of claim 1, wherein:

the processor generates the user-initiated actions to perform an operation requested by the user;

the operation includes at least one of;

(a) purchasing a product;

(b) searching a database;

(c) accessing a web service;

(d) accessing a web page;

(e) performing a business process; and

(d) operating a device; and

the generating of the plurality of user-initiated actions includes digitizing and recognizing words in the spoken dialog under control of the processor.

4. The method of claim 3, wherein the device is one of an mp3 player, a CD player, and a stereo.

5. The method of claim 1, further comprising storing in a memory an action-state table having an action state corresponding to each of the plan sessions, wherein:

the action state indicates at least one of;

(a) whether a corresponding plan session is waiting for a particular user-initiated action; and

(b) a type of user-initiated action for which the corresponding plan session is waiting; and

the distributing of at least one of the user-initiated actions to the plan sessions includes determining from the action-state table a subset of the plan sessions that are waiting for particular user-initiated actions.

6. The method of claim 1, further comprising storing in a memory an activity-level list having an activity level corresponding to the degree of user-interaction activity corresponding to each of the plan sessions,

wherein:

at least one of;

(a) a particular plan session associated with a most-recent query-data output to the user is assigned a highest activity level; and

(b) a plan session associated with a most-recent data output to the user is assigned the highest activity level; and

the distributing of at least one of the user-initiated actions to the plan sessions includes determining from the activity-level list a plan session having the highest activity level.

7. The method of claim 1, further comprising:

upon determining that at least one of the generated user-initiated actions does not correspond to any of the plan sessions executed under control of the processor, initiating a new plan session executed under control of the processor; and

distributing the at least one of the generated user-initiated actions to the newly-initiated plan session.

8. The method of claim 1, further comprising:

outputting data to the user via the user interface from at least one of the plan sessions, wherein the data includes at least one of;

query data output to the user to request response data from the user; and

non-query data output to the user.

9. The method of claim 1, further comprising:

storing a list of plan sessions that have requested to output data to the user;

assigning an output-enabling token to a first of the plan sessions in the list of plan sessions;

where the first plan session is required to wait to receive response data in response to outputting of data by the first plan session, upon determining that the outputting of the data and the waiting to receive response data have been completed by the first plan session, assigning the output-enabling token to a second of the plan sessions in the list of plan sessions; and

where the first plan session is not required to wait to receive response data in response to the outputting of data by the first plan session, upon determining that the outputting of the data has been completed by the first plan session, assigning the output-enabling token to the second of the plan sessions in the list of plan sessions.

10. The method of claim 1, further comprising:

receiving a request from the user via the user interface to at least one of;

(a) pause a particular plan session; and

(b) stop a particular plan session; and

in response to receiving the request, storing in an action-state table an indication of at least one of;

(a) whether the particular plan session is paused; and

(b) whether the particular plan session is stopped.

11. The method of claim 1, further comprising at least one of:

while executing a first of the plan sessions, generating and executing a new plan session that is a copy of at least a portion of the first plan session; and

merging at least two of the plan sessions into a single merged plan session to execute a same portion of plan scripts associated with both of the at least two plan sessions.

12. A hardware-implemented computer-readable storage medium comprising computer program instructions, which, when executed by a processor, cause the processor to perform a method, the method comprising:

receiving spoken dialog from a user at a user interface;

generating a plurality of user-initiated actions by a processor based on the received spoken dialog; and

distributing the plurality of user-initiated actions by the processor to a plurality of plan sessions configured to simultaneously execute under control of the processor, each plan session configured to;

store in a memory a queue of distributed user-initiated actions; and

execute plan scripts to perform the distributed user-initiated actions.

13. The computer-readable storage medium of claim 12, wherein a first user-initiated action is distributed to a first plan session, a second user-initiated action is distributed to a second plan session, and the first and second plan sessions simultaneously execute to perform the first and second user-initiated actions.

14. The computer-readable storage medium of claim 12, wherein:

the processor generates the user-initiated actions to perform an operation requested by the user;

the operation includes at least one of;

(a) purchasing a product; and

(b) searching a database; and

the generating of the plurality of user-initiated actions includes digitizing and recognizing words in the spoken dialog under control of the processor.

15. The computer-readable storage medium of claim 12, wherein:

the method further comprises storing in a memory an action-state table having an action state corresponding to each of the plan sessions;

the action state indicates at least one of;

(a) whether a corresponding plan session is waiting for a particular user-initiated action; and

(b) a type of user-initiated action for which the corresponding plan session is waiting; and

the distributing of at least one of the user-initiated actions to the plan sessions includes determining from the action-state table a subset of the plan sessions that are waiting for particular user-initiated actions.

16. The computer-readable storage medium of claim 12, wherein:

the method further comprises storing in a memory an activity-level list having an activity level corresponding to the degree of user-interaction activity corresponding to each of the plan sessions;

at least one of;

(a) a particular plan session associated with a most-recent query-data output to the user is assigned a highest activity level; and

(b) a plan session associated with a most-recent data output to the user is assigned the highest activity level; and

the distributing of at least one of the user-initiated actions to the plan sessions includes determining from the activity-level list a plan session having the highest activity level.

17. The computer-readable storage medium of claim 12, wherein the method further comprises:

upon determining that at least one of the generated user-initiated actions does not correspond to any of the plan sessions executed under control of the processor, initiating a new plan session executed under control of the processor; and

distributing the at least one of the generated user-initiated actions to the newly-initiated plan session.

18. The computer-readable storage medium of claim 12, wherein the method further comprises:

outputting data to the user via the user interface from at least one of the plan sessions, wherein the data includes at least one of;

(a) query data output to the user to request response data from the user; and

(b) non-query data output to the user.

19. The computer-readable storage medium of claim 12, wherein the method further comprises:

storing a list of plan sessions that have requested to output data to the user;

assigning an output-enabling token to a first of the plan sessions in the list of plan sessions;

where the first plan session is required to wait to receive response data in response to outputting of data by the first plan session, upon determining that the outputting of the data and the waiting to receive response data have been completed by the first plan session, assigning the output-enabling token to a second of the plan sessions in the list of plan sessions; and

where the first plan session is not required to wait to receive response data in response to the outputting of data by the first plan session, upon determining that the outputting of the data has been completed by the first plan session, assigning the output-enabling token to the second of the plan sessions in the list of plan sessions.

20. The computer-readable storage medium of claim 12, wherein the method further comprises:

receiving a request from the user via the user interface to at least one of;

(a) pause a particular plan session; and

(b) stop a particular plan session; and

in response to receiving the request, storing in an action-state table an indication of at least one of;

(a) whether the particular plan session is paused; and

(b) whether the particular plan session is stopped.

21. The computer-readable storage medium of claim 12, wherein the method further comprises at least one of:

while executing a first of the plan sessions, generating and executing a new plan session that is a copy of at least a portion of the first plan session; and

merging at least two of the plan sessions into a single merged plan session to execute a same portion of plan scripts associated with both of the at least two plan sessions.

22. A system, comprising:

a user interface configured to receive spoken dialog from a user; and

a processor configured to;

generate a plurality of user-initiated actions based on the received spoken dialog; and

distribute the plurality of user-initiated actions to a plurality of plan sessions that simultaneously execute under control of the processor, each plan session configured to store in a memory a queue of distributed user-initiated actions and execute plan scripts to perform the distributed user-initiated actions.

23. The system of claim 22, wherein:

the processor is configured to generate the user-initiated actions to perform an operation requested by the user; and

the operation includes at least one of;

(a) purchasing a product; and

(b) searching a database.

24. The system of claim 22, further comprising a memory configured to store an action-state table having an action state corresponding to each of the plan sessions, wherein:

the action state indicates at least one of;

(a) whether a corresponding plan session is waiting for a particular user-initiated action; and

(b) a type of user-initiated action for which the corresponding plan session is waiting; and

the distributing of at least one of the user-initiated actions to the plan sessions includes determining from the action-state table a subset of the plan sessions that are waiting for particular user-initiated actions.

25. The system of claim 22, further comprising:

a memory configured to store an activity-level list having an activity level corresponding to the degree of user-interaction activity corresponding to each of the plurality of plan sessions;

wherein;

at least one of;

(a) a particular plan session associated with the most recent query-data output to the user is assigned a highest activity level; and

(b) a plan session associated with the most recent data output to the user is assigned the highest activity level; and

the distributing of at least one of the user-initiated actions to the plan sessions includes determining from the activity-level list a plan session having the highest activity level.

26. The system of claim 22, wherein:

the processor is configured to control outputting of data to the user via the user interface from at least one of the plan sessions; and

the data includes at least one of;

(a) query data output to request response data to the user; and

(b) non-query data output to the user.

27. The system of claim 22, wherein the processor is configured to at least one of:

while executing a first of the plan sessions, generate and execute a new plan session that is a copy of at least a portion of the first plan session; and

merge at least two of the plan sessions into a single merged plan session to execute a same portion of plan scripts associated with both of the at least two plan sessions.