Automated control-schedule acquisition within an intelligent controller

• US 9,020,646 B2
• Filed: 12/06/2013
• Issued: 04/28/2015
• Est. Priority Date: 10/21/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Matsuoka, Yoky, Lee, Eric A., Hales, Steven A., Stefanski, Mark D., Sharan, Rangoli
• Primary Examiner(s): Bahta, Kidest
• Application Number:
  US14/099,853
• Publication Number:
  US 20140195012A1
• Time in Patent Office: 508 Days
• Field of Search: 700/12, 700/17, 700/276, 700/278, 700/277, 707/809, 706/52, 702182-184, 709/250, 709/233
• US Class Current: 700/276
• CPC Class Codes: F24F 11/30 for purposes related to the operation of the system, e.g. for safety or monitoring
• CPC Class Codes: F24F 11/523 for displaying temperature data
• CPC Class Codes: F24F 11/58 using Internet communication
• CPC Class Codes: F24F 11/62 characterised by the type of control or by internal processing, e.g. using fuzzy logic, adaptive control or estimation of values
• CPC Class Codes: F24F 11/64 using pre-stored data
• CPC Class Codes: G05B 13/0265 the criterion being a learning criterion
• CPC Class Codes: G05D 23/1904 variable in time
• CPC Class Codes: G05D 23/1917 using digital means
• CPC Class Codes: H04L 12/2803 Home automation networks
• CPC Class Codes: H04L 12/2825 Reporting to a device located outside the home and the home network access arrangements H04L12/2856; for remote control or remote monitoring of applications H04L67/025; telephonic communication systems adapted for combination with telemetering systems H04M11/002
• CPC Class Codes: H04L 12/40013 Details regarding a bus controller
• CPC Class Codes: H04L 2012/285 Generic home appliances, e.g. refrigerators

Abstract

The current application is directed to intelligent controllers that initially aggressively learn, and then continue, in a steady-state mode, to monitor, learn, and modify one or more control schedules that specify a desired operational behavior of a device, machine, system, or organization controlled by the intelligent controller. An intelligent controller generally acquires one or more initial control schedules through schedule-creation and schedule-modification interfaces or by accessing a default control schedule stored locally or remotely in a memory or mass-storage device. The intelligent controller then proceeds to learn, over time, a desired operational behavior for the device, machine, system, or organization controlled by the intelligent controller based on immediate-control inputs, schedule-modification inputs, and previous and current control schedules, encoding the desired operational behavior in one or more control schedules and/or sub-schedules.

Claims

1. An intelligent controller comprising:

a processor;

a memory;

a control schedule stored in the memory;

a schedule interface;

a control interface; and

instructions stored within the memory that, when executed by the processor, cause the intelligent controller to;

control, during a monitoring period, an environmental system according to the control schedule, wherein the length of the monitoring period is at least one day in length;

receive immediate-control inputs through the control interface during the monitoring period and record the received immediate-control inputs in the memory;

receive schedule changes through the schedule interface during the monitoring period and record the received schedule changes in the memory;

generate, after the monitoring period, an updated monitoring-period schedule based on the recorded immediate-control inputs and the recorded schedule changes recorded in the memory during the monitoring period, and the control schedule;

substitute, after the monitoring period, the updated monitoring-period schedule for a portion of the control schedule corresponding to the monitoring period; and

propagate, after the monitoring period, the updated monitoring-period schedule to additional time periods within the control schedule.

2. The intelligent controller of claim 1

wherein the processor includes one or more of

one or more local processors, and

one or more remote processors;

and

wherein the memory includes one or more of

one or more local electronic memories,

one or more local mass-storage devices,

one or more remote electronic memories, and

one or more remote mass-storage devices.

3. The intelligent controller of claim 1 wherein control schedule stored in the memory includes one or more setpoints, each setpoint associated with a time and one or more parameter values.

4. The intelligent controller of claim 1 wherein an immediate-control input is made through the control interface to specify one or more parameter values to be implemented by one or more intelligent-controller outputs.

5. The intelligent controller of claim 1 wherein a control-schedule change is input through the schedule interface to specify a control-schedule setpoint to be implemented by one or more intelligent-controller outputs when a current time is equal to the time associated with the schedule setpoint.

6. The intelligent controller of claim 1 wherein the instructions generate an updated monitoring-period schedule, after the monitoring period, based on the recorded immediate-control inputs, recorded schedule changes, and the control schedule by:

combining the recorded immediate-control inputs, recorded schedule changes, and setpoints of the control schedule to produce a provisional schedule;

clustering the immediate-control inputs, schedule changes, and setpoints of the provisional schedule into one or more clusters;

resolving the one or more clusters within the provisional schedule; and

generating the updated monitoring-period schedule from the resolved clusters of the provisional schedule.

7. The intelligent controller of claim 6 wherein combining the recorded immediate-control inputs, recorded schedule changes, and setpoints of the control schedule to produce the provisional schedule further comprises:

representing immediate-control inputs as time-associated values of one or more parameters within a portion of the control schedule corresponding to the monitoring period;

representing recorded control-schedule changes as time-associated values of one or more parameters, additionally associated with input times, within the portion of the control schedule corresponding to the monitoring period; and

considering the represented immediate-control inputs, retrospective control-schedule changes, and setpoints of the provisional schedule to be events.

8. The intelligent controller of claim 7 wherein clustering the immediate-control inputs, schedule changes, and setpoints within the provisional schedule into one or more clusters further comprises:

collecting together, as a separate cluster, each group of events that are each within a first threshold time interval of another event of the separate cluster; and

collecting, as a separate cluster, each event that is not within the first threshold time interval of another event.

9. The intelligent controller of claim 8 wherein resolving the one or more clusters within the provisional schedule further includes replacing each cluster of events with one of:

no event, when no control-change trend is discerned in the cluster;

one event, when a single-event control-change trend is discerned in the cluster; and

two events, when a two-event control-change trend is discerned in the cluster.

10. The intelligent controller of claim 1 wherein the instructions propagate the updated monitoring-period schedule to additional time periods within the control schedule by:

selecting, according to one or more rules, one or more additional time periods within the control schedule related to the time period of the updated monitoring-period schedule; and

to each of the selected time periods,

copying setpoints corresponding to immediate-control inputs of the provisional monitoring-period schedule onto a portion of the control schedule corresponding to the selected time period, and

resolving the portion of the control schedule corresponding to the selected time period.

11. The intelligent controller of claim 10 wherein resolving the portion of the control schedule corresponding to the selected time period further comprises:

applying, to each copied setpoint, one or more rules that result in one of

deleting the copied setpoint,

deleting an existing setpoint,

moving either the copied setpoint or another setpoint in time, and

no change to the copied setpoint.

12. The intelligent controller of claim 1 further comprising:

at least three phases of control-schedule learning implemented by instructions stored in the memory, including

an initial phase,

at least one aggressive-learning phase, and

one or more steady-state learning phases.

13. The intelligent controller of claim 12

wherein, during the aggressive-learning phase, learning parameters are provided values to encourage frequent, additional immediate-control inputs;

and

wherein, during the steady-state learning phase, learning parameters are provided values to discourage frequent, additional immediate-control inputs.

14. The intelligent controller of claim 1 wherein the intelligent controller controls one or more of:

an HVAC unit;

a furnace;

an air conditioner;

a heat pump;

an irrigation system;

a pump;

a fan;

one or more light sources;

a machine;

a device;

an organization; and

a system.

15. The intelligent controller of claim 1 wherein the setpoints of the control schedule specify one or more of:

temperature;

flow rate of a liquid or gas;

rate of energy dissipation;

pressure;

current density;

voltage;

a machine setting;

a position of a machine component;

a computational state;

a mechanical state; and

process throughput.

16. An intelligent controller comprising:

a processor;

a memory;

a schedule interface;

a control interface; and

instructions stored within the memory that, when executed by the processor,

acquire an initial control schedule and store the initial control schedule in memory as a current control schedule,

during an aggressive-learning phase, modify the current control schedule based on immediate-control inputs received through the control interface and schedule changes received through the schedule interface, and

during a steady-state learning phase that follows the aggressive-learning phase, modify the current control schedule based on immediate-control inputs received through the control interface, schedule changes received through the schedule interface, and historical control schedules.

17. The intelligent controller of claim 16 wherein, during the aggressive-learning phase and the steady-state learning phase, the intelligent controller

receives immediate-control inputs through the control interface during a monitoring period and records the received immediate-control inputs in memory;

receives schedule changes through the schedule interface during the monitoring period and records received schedule changes of at least one type in the memory;

generates an updated monitoring-period schedule, after the monitoring period, based on the recorded immediate-control inputs, recorded schedule changes, and the current control schedule;

substitutes the updated monitoring-period schedule for a portion of the current control schedule corresponding to the monitoring period; and

propagates the updated monitoring-period schedule to additional time periods within the current control schedule.

18. The intelligent controller of claim 17 wherein the intelligent controller generates an updated monitoring-period schedule, after the monitoring period, based on the recorded immediate-control inputs, recorded schedule changes, and the current control schedule by:

combining the recorded immediate-control inputs, recorded schedule changes, and setpoints of the current control schedule to produce a provisional schedule;

clustering the immediate-control inputs, schedule changes, and setpoints of the provisional schedule into one or more clusters;

resolving the one or more clusters within the provisional schedule; and

generating the updated monitoring-period schedule from the resolved clusters of the provisional schedule.

19. A method that automatically learns a control schedule implemented within an intelligent controller that includes one or more processors and one or more memories, the method comprising:

acquiring an initial control schedule and storing the initial control schedule in one of the one or more memories as a current control schedule;

during each of a set of successive monitoring periods;

controlling, during the monitoring period, an environmental system according to the current control schedule, wherein the length of the monitoring period is at least one day in length;

receiving immediate-control inputs through a control interface and recording the received immediate-control inputs in one of the one or more memories, and

receiving schedule changes through a schedule interface and recording the received schedule changes of at least one type in the memory; and

after each of the set of successive monitoring periods;

generating an updated monitoring-period schedule based on the recorded immediate-control inputs, the recorded schedule changes, and the current control schedule,

substituting the updated monitoring-period schedule for a portion of the current control schedule corresponding to the monitoring period, and

propagating the updated monitoring-period schedule to additional time periods within the current control schedule.

20. The method of claim 19 further including:

acquiring an initial control schedule and storing the initial control schedule in one of the one or more memories as a current control schedule,

during an aggressive-learning phase, modifying the current control schedule based on immediate-control inputs received through a control interface and schedule changes received through a schedule interface, and

during a steady-state learning phase, modifying the current control schedule based on immediate-control inputs received through the control interface, schedule changes received through the schedule interface, and historical control schedules.