Heuristic method of automated and learning control, and building automation systems thereof
AssigneePassiveLogic, Inc.
InventorTroy Aaron Harvey
Life of this patent
Estimating the term…
Abstract
Apparatuses, systems, and methods of physical-model based building automation using in-situ regression to optimize control systems are presented. A simulation engine is configured to simulate a behavior or a controlled system using a physical model for the controlled system. A data stream comprises data from a controlled system. A training loop is configured to compare an output of a simulation engine to a data stream using a heuristic so that a physical model is regressed in a manner that the output of the simulation engine approaches the data stream.
Claims20
A building system controller comprising:
- a controlled system;
- a physical model for the controlled system, the physical model comprising a mathematical model of the controlled system;
- a simulation engine configured to simulate a behavior of the controlled system using the physical model for the controlled system;
- a heuristic;
- a data stream, comprising data from the controlled system;
- and
- a training loop configured to compare an output of the simulation engine to the data stream using the heuristic, such that the physical model is regressed in a manner that the output of the simulation engine approaches the data stream.
A system comprising:
- a coupled building system;
- a data stream;
- at least one memory;
- and
- one or more processors operatively coupled to the data stream and the one or more processors operatively coupled to the at least one memory, wherein the one or more processors are configured to execute program code stored in the at least one memory to:
- control the coupled building system;
- receive sensor measurements obtained from the data stream;
- store the sensor measurements in the at least one memory;
- compute a mathematical model that substantially imitates a behavior of the controlled system, generating a predicted data stream measurement from the mathematical model;
- analyze a variation between the stored sensor measurements and the predicted data stream measurement, using a cost function;
- iterate a process of computation and analysis, by changing parameters of the mathematical model to regress the mathematical model such that a subsequently generated predicted data stream measurement from the regressed mathematical model approaches a subsequently received sensor measurement from the data stream;
- and
- control the coupled building system using predicted actuation from the regressed mathematical model.
The system of claim 21, wherein the mathematical model is computed repeatedly, in arbitrary order, to evaluate each of an arbitrary number of different physical models, the one or more processors comprise one or more computational entities in a network of simulation engines, and the one or more processors comprise collocated computational entities as well as remote computational entities.
Dependencies are inferred from the claim text.