Building management system with artificial intelligence for unified agent based control of building subsystems

Patent metadata

Field	Value
Publication number	US-10901373-B2
Application number	US-201816008885-A
Country	US
Kind code	B2
Filing date	Jun 14, 2018
Priority date	Jun 15, 2017
Publication date	Jan 26, 2021
Grant date	Jan 26, 2021

Abstract

A building management system with artificial intelligence based control of a building includes data collectors are configured to receive data and generate data streams for subsystems of the building. The system includes a learning engine configured to identify a building state of the building by correlating data of the data streams for the subsystems and provide the identified building state to cognitive agents. The system includes the cognitive agents, each of the cognitive agents configured to receive the identified building state from the learning engine, generate a control decision based on the received building state, and operate at least one of the plurality of subsystems of the building to control a physical condition of the building based on the control decision.

First claim

What is claimed is:

1. A building management system with artificial intelligence based control of a building, the system comprising:
a processing system comprising one or more processing circuits and one or more non-transitory memory devices, wherein the processing system is configured to implement a plurality of data collectors, a learning engine, and a plurality of cognitive agents:
wherein the plurality of data collectors are configured to:
receive data from a plurality of subsystems of the building; and
generate a plurality of data streams based on the data, wherein a data stream of the plurality of data streams is associated with each of the plurality of subsystems of the building;

wherein the learning engine is configured to:
identify a building event occurring at the building by correlating data of the plurality of data streams for the plurality of subsystems; and
provide the building event to the plurality of cognitive agents;

wherein the plurality of cognitive agents comprise a first cognitive agent configured to operate a first building subsystem of a first type, wherein the plurality of cognitive agents comprise a second cognitive agent configured to operate a second building subsystem of a second type different than the first type, wherein operation of the first building subsystem by the first cognitive agent and operation of the second building subsystem by the second cognitive agent together achieve a common outcome responsive to the building event.

2. The system of claim 1 , wherein each of the plurality of cognitive agents are configured to generate a control decision based on the building event by generating the control decision with a reactive function, wherein the reactive function is unique to a type of building subsystem that each of the plurality of cognitive agents is associated with.

3. The system of claim 1 , wherein the building event of the building is an emergency event indicating that the building is experiencing an emergency.

4. The system of claim 1 , wherein the learning engine is configured to identify the building event of the building by correlating data from the plurality of data streams for the plurality of subsystems by:
extracting one or more features from the plurality of data streams, wherein the one or more features comprises a portion of the plurality of data streams;
transforming the one or more features; and
identifying the building event by correlating the one or more features by performing machine learning on the one or more features.

5. The system of claim 1 , wherein the plurality of data streams comprises an augmented reality data stream of an augmented reality system;
wherein the learning engine is configured to:
identify a building component of interest to user based on the augmented reality data stream;
extract one or more features from the plurality of data streams of the plurality of subsystems, the one or more features being associated with the building component of interest;
determine, based on the one or more features, a reason that the building component of interest is of interest to the user; and
provide the reason to the augmented reality system, wherein the augmented reality system is configured to present the reason to the user.

6. The system of claim 1 , wherein the processing system is configured to implement a prioritization engine, wherein the prioritization engine is configured to:
generate a prioritization function for each of the plurality of cognitive agents; and
send the prioritization function of each of the plurality of cognitive agents to the plurality of cognitive agents;
wherein the plurality of cognitive agents are configured to determine whether to override an operation of a particular subsystem of the plurality of subsystems based on the prioritization function and data of the particular subsystem or the building event.

7. The system of claim 1 , wherein the plurality of data streams comprise a natural language data stream of a natural language input system;
wherein the learning engine is configured to:
determine, based on the natural language data stream, a particular event state of interest to the user, wherein the particular event of interest comprises a building draft event indicating that the building is experiencing an air draft; and
provide the building draft event to each of the plurality of cognitive agents.

8. The system of claim 7 , wherein one of the plurality of cognitive agents are associated with a security subsystem of the plurality of subsystems, wherein the one of the plurality of cognitive agents associated with the security subsystem is configured to determine, based on data of the security subsystem, whether one or more doors or windows are open in response to a reception of the building draft event; and
wherein a second one of the plurality of cognitive agents is associated with an HVAC subsystem of the plurality of subsystems, wherein the second one of the plurality of cognitive agents associated with the HVAC subsystem is configured to determine, based on data of the HVAC subsystem, whether a damper of the HVAC subsystem is stuck in response to receiving the building draft state.

9. The system of claim 1 , wherein the plurality of subsystems comprise a fire detection subsystem, a video surveillance subsystem, and a heating, ventilation, and air conditioning (HVAC) subsystem;
wherein the learning engine is configured to identify the building event of the building by correlating data of a first data stream of the fire detection subsystem, a second data stream of the video surveillance subsystem, and a third data stream of the HVAC subsystem, wherein the building event is a fire event indicative of a fire in the building;
wherein each of the plurality of cognitive agents is configured to control the fire in the building in response to a reception of the fire event.

10. The system of claim 9 , wherein each of the plurality of cognitive agents is configured to generate a particular control decision based on the fire event by determining the particular control decision based on a plurality of reactive functions, the plurality of reactive functions comprising a reactive function for each of the plurality of cognitive agents, wherein the plurality of reactive functions comprise:
a first reactive function that generates the particular control decision for the HVAC subsystem, the particular control decision for the HVAC subsystem causing the HVAC subsystem to operate to reduce an oxygen level of an unoccupied area of the building;
a second reactive function that generates the particular control decision for the HVAC subsystem, the particular control decision for the HVAC subsystem causing the HVAC subsystem to operate to exhaust smoke from exit routes of the building; and
a third reactive function that generates the particular control decision for a human-machine interaction system, the particular control decision for the human-machine interaction system causing the human-machine interaction system to provide occupants of the occupied area with details regarding the exit routes.

11. The system of claim 9 , wherein the building state is a fire event, wherein one of the plurality of cognitive agents is associated with a plurality of dampers of the HVAC subsystem, wherein the one of the plurality of cognitive agents is configured to:
receive a humidity value of outdoor air outside the building, wherein the humidity value is measured by the HVAC subsystem;
determine whether the humidity value of outdoor air is greater than a predefined threshold;
generate a control decision to open dampers in an area of the building associated with the fire in response to a determinati

Assignees

• Johnson Controls Tech Co

Inventors

• Locke Robert
• Sinha Sudhi R

Classifications

• G05B13/0265 Primary

  the criterion being a learning criterion · CPC title

• H04L12/2827 Primary

  Reporting to a device within the home network; wherein the reception of the information reported automatically triggers the execution of a home appliance functionality · CPC title

• G05B15/02

  electric · CPC title

• H04L41/046

  comprising network management agents or mobile agents therefor · CPC title

• G06T19/006

  Mixed reality (object pose determination, tracking or camera calibration for mixed reality G06T7/00) · CPC title