GENETIC LEARNING FOR ENVIRONMENTAL CONTROL AUTOMATION

• US 20140129032A1
• Filed: 11/06/2013
• Published: 05/08/2014
• Est. Priority Date: 11/07/2012
• Status: Abandoned Application

• Current Assignee: Think Automatic, LLC
• Original Assignee: Think Automatic, LLC
• Inventors: HARRIS, Stephen
• Application Number:
  US14/073,695
• Publication Number:
  US 20140129032A1
• Time in Patent Office: Days
• US Class Current: 700/275
• CPC Class Codes: G05B 13/0265 the criterion being a learning criterion
• CPC Class Codes: G05B 15/02 electric
• CPC Class Codes: G05B 2219/2642 Domotique, domestic, home control, automation, smart house
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

Disclosed is a method and apparatus for an environmental control system in which a genetic learning algorithm creates scenes and scene triggers and in which a fitness function scores the scenes through end-user interaction.

Claims

1. A computer implement method of controlling an environmental control device, the method comprising:

at a first computer, receiving a first device report from a first environmental control device, which device report comprises at least one of event and state information for the first environmental control device;

sorting the device report by location and environmental control device;

defining a first scene comprising a first combination of device events and device status and a first scene trigger;

defining a second scene comprising a second combination of device events and device status and a second scene trigger;

creating a first scene trigger group comprising scenes associated with the first and second scene triggers; and

utilizing a fitness function to determining a scene trigger score for the scenes in the first scene trigger group.

2. The method according to claim 1, further comprising:

determining that the first or a subsequent device report comprises a scene trigger in the first scene trigger group;

selecting a scene in the first scene trigger group with the highest scene trigger score; and

implementing the selected scene with the highest scene trigger score.

3. The method according to claim 2, further comprising starting a countdown clock upon determining that the first or the subsequent device report comprises a scene trigger in the first scene trigger group.

4. The method according to claim 3, while the countdown clock has not elapsed, receiving a second device report comprising a scene trigger in the first scene trigger group and determining whether any scene candidates remain in the first scene trigger group.

5. The method according to claim 4, further comprising determining that at least one scene candidate remains in the first scene trigger group, selecting a scene in the first scene trigger group with the next highest scene trigger score, and implementing the selected scene with the next highest scene trigger score.

6. The method according to claim 4, further comprising determining that no scene candidate remain in the first scene trigger group, executing a genetic operator to generate new scene candidates, and including the new scene candidates in the first scene trigger group.

7. The method according to claim 6, further comprising utilizing the fitness function to determine scene trigger scores for the new scene candidates, and selecting the next scene candidate in the first scene trigger group which has the highest scene trigger score.

8. The method according to claim 6, wherein the genetic operator comprises selecting the scenes in the first scene trigger group with the highest scene trigger score, cross-combining the device commands for each device in each scene, and saving the result as scene candidates.

9. The method according to claim 6, wherein the genetic operator comprises selecting the scenes in the first scene trigger group with the highest scene trigger score, randomly combining the device commands in each scene, and saving the result as scene candidates.

10. The method according to claim 1, wherein the fitness function groups scenes by location and scene trigger, determines the length of time each scene was active for, and increments or decrements the scene trigger score for the scene if the length of time the scene was active for is greater or less than a temporal threshold.

11. The method according to claim 1, further comprising assigning a scene identifier to combinations of device events and device status which have not previously been received in a device report.

12. The method according to claim 1, further comprising assigning an existing scene identifier to combinations of device events and device status which have previously been received in a device report.

13. The method according to claim 1, wherein at least one of the first or second device reports comprises event and state information input into the environmental control device by a user.

14. The method according to claim 1, wherein at least one of the first or second device report comprises event and state information input into the environmental control device by an environmental control device controller.

15. The method according to claim 1, wherein the first and second scene triggers are the same and the first scene trigger group contains only scene triggers which are the same.

16. A non-transient computer-readable medium having stored thereon instructions that, when executed by a processor, configure the processor to:

receive a first device report from a first environmental control device, which device report comprises at least one of event and state information for the first environmental control device;

sort the device report by location and environmental control device;

define a first scene comprising a first combination of device events and device status and a first scene trigger;

define a second scene comprising a second combination of device events and device status and a second scene trigger;

create a first scene trigger group comprising scenes associated with the first and second scene triggers; and

utilize a fitness function to determining a scene trigger score for the scenes in the first scene trigger group.

17. A computer implement method of registering environmental control devices for use in an environmental control system, the method comprising:

at a first computer, receiving a first set of device commands and attributes for a first environmental control device;

assigning a first device identifier to the first environmental control device;

receiving a second set of device commands and attributes for a second environmental control device;

assigning a second device identifier to the second environmental control device;

wherein the first and second set of attributes comprise a common location;

for environmental control devices in the common location, performing a genetic operator on the sets device commands to generate scene candidates; and

performing a fitness function on the scene candidates to develop a scene trigger score.

18. The method according to claim 17, wherein the genetic operator comprises cross-combining the device commands available to each environmental control device.

19. The method according to claim 17, wherein the genetic operator comprises randomly combining the device commands available to each environmental control device.

20. The method according to claim 17, wherein the fitness function comprises assigning a default scene trigger score or the average scene trigger score for similar scenes in locations other than the common location.