System and method for controlling an operation of a device

Patent metadata

Field	Value
Publication number	US-12353176-B2
Application number	US-202217819557-A
Country	US
Kind code	B2
Filing date	Aug 12, 2022
Priority date	Aug 12, 2022
Publication date	Jul 8, 2025
Grant date	Jul 8, 2025

Abstract

The present disclosure provides a feedback controller and method for controlling an operation of a device at different control steps. The feedback controller comprises at least one processor, and the memory having instructions stored thereon that, when executed by the at least one processor, causes the feedback controller, for a control step, to collect a measurement indicative of a state of the device at the control step, and execute, recursively until a termination condition is met, a probabilistic solver parameterized on a control input to an actuator operating the device to produce a control input for the control step. The feedback controller is further configured to control the actuator operating the device based on the produced control input.

First claim

We claim:

1. A feedback controller for controlling an operation of a device at different control steps based on a feedback signal including measurements indicative of a state of the device at the different control steps, the feedback controller comprising: at least one processor; and memory having instructions stored thereon that, when executed by the at least one processor, causes the feedback controller, for a control step of the different control steps, to:
collect a measurement indicative of the state of the device at the control step;
execute, recursively until a termination condition is met, a probabilistic solver parameterized on a control input to an actuator operating the device to produce the control input for the control step, wherein, during each of the executions, the probabilistic solver is configured to
estimate a probabilistic distribution function (PDF) of predicted values of the control input from a PDF of values of the control input using a prediction model;
evaluate a cost function of an optimal control problem for controlling the device based on a simulation of the operation of the device with the measurement and a value of the control input sampled from the PDF of the predicted values of the control input to produce a performance metric of the operation of the device; and
estimate a PDF of simulated values connected to the control input based on a measurement model connecting the performance metric of the operation of the device with the control input; and
correct the PDF of the predicted values of the control input based on the PDF of the simulated values connected to the control input to produce the PDF of the values of the control input; and

control the actuator operating the device using at least a mean of the PDF of the values of the control input.

2. The feedback controller of claim 1 , wherein the prediction model is an identity model.

3. The feedback controller of claim 1 , wherein the prediction model is given by a gradient of the cost function with respect to the control input.

4. The feedback controller of claim 1 , wherein, for each iteration of iterations performed for each of the control steps, the probabilistic solver is further configured to evaluate the cost function for multiple samples defining simulation trials of the PDF of the predicted values.

5. The feedback controller of claim 4 , wherein to produce the PDF of the simulated values connected to the control input from multiple performance metrics, the probabilistic solver is further configured to determine a weighted combination of evaluations of the multiple samples of the PDF of the predicted values.

6. The feedback controller of claim 4 , wherein the cost function for the multiple samples of the PDF of the predicted values is evaluated in parallel using multiple processors.

7. The controller of claim 1 , wherein the termination condition is based on a similarity metric between the PDF of values of the control input and a PDF of the values of the control input at a previous iteration.

8. The feedback controller of claim 1 , wherein the probabilistic solver is a Kalman filter.

9. The feedback controller of claim 1 , wherein the probabilistic solver is an unscented Kalman filter performing a gradient-free correction of the predicted values of the control input.

10. The feedback controller of claim 9 , wherein to update the PDF of the predicted values using the gradient-free correction, the unscented Kalman filter is configured to produce the PDF of the simulated values connected to the control input using evaluations of the cost function.

11. The feedback controller of claim 9 , wherein the unscented Kalman filter evaluates the cost function for multiple sigma points of the PDF of the predicted values.

12. The feedback controller of claim 11 , wherein, to determine the sigma points, the unscented Kalman filter is configured to generate the sigma points based on a predicted mean of the PDF of the predicted values and a covariance matrix of the PDF of the predicted values.

13. The feedback controller of claim 12 , wherein the sigma points are determined based on a Cholesky decomposition.

14. A method for controlling an operation of a device at different control steps based on a feedback signal including measurements indicative of a state of the device at the different control steps, the method comprising:
collecting a measurement indicative of the state of the device at a control step of the different control steps;
executing, recursively until a termination condition is met, a probabilistic solver parameterized on a control input to an actuator operating the device to produce the control input for the control step, wherein, during each of the executions, the probabilistic solver is configured to
estimate a probabilistic distribution function (PDF) of predicted values of the control input from a PDF of values of the control input using a prediction model;
evaluate a cost function of an optimal control problem for controlling the device based on a simulation of the operation of the device with the measurement and a value of the control input sampled from the PDF of the predicted values of the control input to produce a performance metric of the operation of the device; and
estimate a PDF of simulated values connected to the control input based on a measurement model connecting the performance metric of the operation of the device with the control input; and
correct the PDF of the predicted values of the control input based on the PDF of the simulated values connected to the control input to produce the PDF of the values of the control input; and

controlling the actuator operating the device using at least a mean of the PDF of the values of the control input.

15. The method of claim 14 , wherein the prediction model is an identity model.

16. The method of claim 14 , wherein the prediction model is given by a gradient of the cost function with respect to the control input.

17. The method of claim 14 , wherein, for each iteration of iterations performed for each of the control steps, the probabilistic solver is further configured to evaluate the cost function for multiple samples defining simulation trials of the PDF of the predicted values.

18. The method of claim 17 , wherein to produce the PDF of the simulated values connected to the control input from multiple performance metrics, the probabilistic solver is further configured to determine a weighted combination of evaluations of the multiple samples of the PDF of the predicted values.

19. The method of claim 14 , wherein the termination condition is based on a similarity metric between the PDF of the values of the control input and a PDF of values of the control input at a previous iteration.

20. A non-transitory computer-readable storage medium embodied thereon a program executable by a processor for performing a method for controlling an operation of a device at different control steps based on a feedback signal including measurements indicative of a state of the device at the different control steps, the method comprising:
collecting a measurement indicative of the state of the device at a control step of the different control steps;
executing, recursively until a termination condition is met, a probabilistic solver parameterized on a control input to an actuator operating the device to produce the control input for the control step, wherein, during each of the executions, the probabilistic solver is configured to
estimate a probabilistic distribution function (PD

Assignees

• Mitsubishi Electric Res Laboratories Inc

Inventors

• Menner Marcel
• Di Cairano Stefano
• Berntorp Karl
• Chakrabarty Ankush

Classifications

• G05B13/041

  in which a variable is automatically adjusted to optimise the performance · CPC title

• G05B6/02

  electric · CPC title

• G05B2219/42061

  Stochastic predictive controller spc · CPC title

• G05B2219/41146

  Kalman filter · CPC title

• G05B13/048 Primary

  using a predictor · CPC title