CALIBRATING REAL-WORLD SYSTEMS USING SIMULATION LEARNING

1 . A method comprising:

receiving simulator input for one or more simulator models corresponding to one or more aspects of an engine;

simulating, based at least on one or more simulator models operating on the simulator input, operation of the engine; and

generating, based at least on simulator output from simulating operation of the engine, calibration data corresponding to one or more electronically controllable components of the engine.

2 . The method of claim 1 , wherein receiving simulator input further comprises:

receiving vehicle operation data based on non-synthetic data that includes sample data from measuring operation of a physical engine being modeled by the one or more simulator models, sample data from engines that are similar, but distinct, from the engine being simulated, or data specified by subject matter experts; and

receiving synthetic data based on a data-driven generation of simulation input.

3 . The method of claim 2 , further comprising:

generating the data-driven simulation input based on combinations of possible simulator inputs, wherein generating the data-driven simulation input is not based on specifications for a given objective for engine operation.

4 . The method of claim 2 , wherein the vehicle operation data includes operation sensor data based on physical operation of the engine and includes operator control inputs based on vehicle controls operated during physical operation of the engine.

5 . The method of claim 1 , wherein generating the calibration data further comprises:

training, based on a learning algorithm operating on the simulator input, simulator output, and one or more target objectives, one or more trained models.

6 . The method of claim 5 , wherein the calibration data includes the one or more trained models, and wherein the simulator input includes: sensor data collected from a physical engine being simulated, sensor data from a physical engine similar to, but distinct from, the engine being simulated, or sensor data from a computing device independent of the engine being simulated that includes driving conditions, environmental conditions, or sensor data from a computing device independent from the vehicle.

7 . The method of claim 6 , wherein the one or more trained models include a valve control model to control an electronically controllable valve.

8 . A computer program product disposed upon a computer-readable medium, the computer program product comprising computer program instructions that, when executed, cause a computer to carry out the steps of:

receiving simulator input for one or more simulator models corresponding to one or more aspects of an engine;

simulating, based at least on one or more simulator models operating on the simulator input, operation of the engine; and

generating, based at least on simulator output from simulating operation of the engine, calibration data corresponding to one or more electronically controllable components of the engine.

9 . The computer program product of claim 8 , wherein receiving simulator input for one or more simulator models corresponding to one or more aspects of an engine comprises:

receiving vehicle operation data based on sample data from operation of a physical engine being modeled by the one or more simulator models; and

receiving synthetic data based on a data-driven generation of simulation input.

10 . The computer program product of claim 9 , wherein the computer program instructions, when executed, further cause the computer to carry out the steps of:

generating the data-driven simulation input based on combinations of possible simulator inputs, wherein generating the data-driven simulation input is not based on specifications for a given objective for engine operation.

11 . The computer program product of claim 9 , wherein the vehicle operation data includes operation sensor data based on physical operation of the engine and includes operator control inputs based on vehicle controls operated during physical operation of the engine.

12 . The computer program product of claim 8 , wherein generating the calibration data further comprises:

training, based on a learning algorithm operating on the simulator input, simulator output, and one or more target objectives, one or more trained models.

13 . The computer program product of claim 12 , wherein the calibration data includes the one or more trained models.

14 . The computer program product of claim 13 , wherein the one or more trained models include a valve control model to control an electronically controllable valve.

15 . An apparatus comprising a computer processor, a computer memory operatively coupled to the computer processor, the computer memory having disposed within it computer program instructions that, when executed by the computer processor, cause the apparatus to carry out the steps of:

receiving simulator input for one or more simulator models corresponding to one or more aspects of an engine;

simulating, based at least on one or more simulator models operating on the simulator input, operation of the engine; and

generating, based at least on simulator output from simulating operation of the engine, calibration data corresponding to one or more electronically controllable components of the engine.

16 . The apparatus of claim 15 , wherein receiving simulator input further comprises:

receiving vehicle operation data based on sample data from operation of a physical engine being modeled by the one or more simulator models; and

receiving synthetic data based on a data-driven generation of simulation input.

17 . The apparatus of claim 16 , wherein the computer program instructions, when executed by the computer processor, further cause the apparatus to carry out the steps of:

generating the data-driven simulation input based on combinations of possible simulator inputs, wherein generating the data-driven simulation input is not based on specifications for a given objective for engine operation.

18 . The apparatus of claim 16 , wherein the vehicle operation data includes operation sensor data based on physical operation of the engine and includes operator control inputs based on vehicle controls operated during physical operation of the engine.

19 . The apparatus of claim 15 , wherein generating the calibration data further comprises:

training, based on a learning algorithm operating on the simulator input, simulator output, and one or more target objectives, one or more trained models.

20 . The apparatus of claim 19 , wherein the calibration data includes the one or more trained models.