System and method for calibrating feedback controllers

Patent metadata

Field	Value
Publication number	US-11977374-B2
Application number	US-202117403222-A
Country	US
Kind code	B2
Filing date	Aug 16, 2021
Priority date	Aug 2, 2021
Publication date	May 7, 2024
Grant date	May 7, 2024

Abstract

A system for controlling an operation of a machine for performing a task is disclosed. The system submits a sequence of control inputs to the machine and receives a feedback signal. The system further determines, at each control step, a current control input for controlling the machine based on the feedback signal including a current measurement of a current state of the system by applying a control policy transforming the current measurement into the current control input based on current values of control parameters in a set of control parameters of a feedback controller. Furthermore, the system may iteratively update a state of the feedback controller defined by the control parameters using a prediction model predicting values of the control parameters and a measurement model updating the predicted values to produce the current values of the control parameters that explain the sequence of measurements according to a performance objective.

First claim

The invention claimed is:

1. A system for controlling an operation of a machine for performing a task, comprising:
a transceiver configured to submit a sequence of control inputs to the machine and to receive a feedback signal including a corresponding sequence of measurements, wherein each measurement is indicative of a state of the machine caused by the corresponding control input;
a feedback controller configured to determine, at each control step, a current control input for controlling the machine based on the feedback signal including a current measurement of a current state of the machine by applying a control policy transforming the current measurement into the current control input based on current values of control parameters in a set of control parameters of the feedback controller to achieve a performance objective of the feedback controller, wherein the performance objective comprises a cost function defining a deviation of the state of the machine from a reference state of the machine; and
a Kalman filter configured to iteratively update a state of the feedback controller defined by the control parameters using a prediction model for predicting values of the control parameters subject to process noise, and a measurement model for updating the predicted values of the control parameters based on the sequence of measurements subject to measurement noise, to produce the current values of the control parameters that are associated with the sequence of measurements according to the performance objective.

2. The system of claim 1 , wherein the Kalman filter is further configured to adjust a Kalman gain for calibrating multiple interdependent control parameters.

3. The system of claim 1 , wherein the prediction model is an identity model that is configured to predict that the control parameters remain fixed within a variance defined by the process noise.

4. The system of claim 1 , wherein the prediction model is configured to predict at least some control parameters within a variance defined by the process noise based on an algebraic relationship with the state of the machine.

5. The system of claim 1 , wherein the performance objective is different from the control policy.

6. The system of claim 1 , wherein the measurement model is configured to update the control parameters by optimizing the cost function.

7. The system of claim 1 , wherein the performance objective further comprises one or a combination of a cost function for the state exceeding an optimal region of operation, a cost function if a reference state is overshot by a certain value, a cost function for oscillations of the state, and a cost function if the state changes between time steps.

8. The system of claim 1 , wherein the control parameters include one or a combination of: (i) one or multiple gains of the feedback controller, (ii) one or multiple structural parameters of the machine, (iii) one or multiple coefficients of one or multiple filters used by the feedback controller, or (iv) one or multiple weights of a neural-network controller.

9. The system of claim 1 , wherein to produce the control parameters, the Kalman filter is configured to update coefficients of a basis function in one or multiple state-dependent regions.

10. The system of claim 1 , wherein to produce the control parameters, the Kalman filter is configured to update coefficients of a basis function in multiple state-dependent regions along with a boundary that separates the multiple state-dependent regions.

11. The system of claim 1 , wherein the Kalman filter is an extended Kalman filter (EKF) that is configured to compute a Kalman gain by computing a gradient of the performance objective.

12. The system of claim 1 , wherein the Kalman filter is an unscented Kalman filter (UKF) that is configured to compute a Kalman gain by evaluating the control parameters with respect to the performance objective.

13. The system of claim 1 , wherein the feedback controller is configured to determine the current control input subject to a constraint on the operation of the machine thereby handling the discontinuity of the control.

14. The system of claim 1 , further comprising a safety check module configured to:
perform a check associated with whether the values of the control parameters produced by the Kalman filter satisfy a safety check according to the control policy; and
update the control parameters of the feedback controller with the control parameters produced by the Kalman filter when the safety check is satisfied.

15. The system of claim 14 , wherein when the values of the control parameters produced by the Kalman filter do not satisfy the safety check, the Kalman filter is further configured to iteratively produce new values of the control parameters until the safety check is satisfied.

16. The system of claim 14 , wherein the safety check includes one or a combination of a boundedness of the state to an origin and a decreasing cost of the state.

17. A method for controlling an operation of a machine for performing a task, comprising:
submitting a sequence of control inputs to the machine;
receiving a feedback signal including a corresponding sequence of measurements, wherein each measurement is indicative of a state of the machine caused by the corresponding control input;
determining, at each control step, a current control input for controlling the machine based on the feedback signal including a current measurement of a current state of the machine by applying a control policy transforming the current measurement into the current control input based on current values of control parameters in a set of control parameters of a feedback controller to achieve a performance objective of the feedback controller, wherein the performance objective comprises a cost function defining a deviation of the state of the machine from a reference state of the machine; and
iteratively updating a state of the feedback controller defined by the control parameters using a prediction model predicting values of the control parameters subject to process noise and a measurement model updating the predicted values of the control parameters based on the sequence of measurements subject to measurement noise to produce the current values of the control parameters that explain the sequence of measurements according to the performance objective.

18. 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 machine for performing a task, the method comprising:
submitting a sequence of control inputs to the machine;
receiving a feedback signal including a corresponding sequence of measurements, wherein each measurement is indicative of a state of the machine caused by the corresponding control input;
determining, at each control step, a current control input for controlling the machine based on the feedback signal including a current measurement of a current state of the machine by applying a control policy transforming the current measurement into the current control input based on current values of control parameters in a set of control parameters of a feedback controller to achieve a performance objective of the feedback controller, wherein the performance objective comprises a cost function defining a deviation of the state of the machine from a reference state of the machine; and
iteratively updating a state of the feedback controller defined by the control parameters using a prediction model predicting values of the control parameters subj

Assignees

• Mitsubishi Electric Res Laboratories Inc

Inventors

• Menner Marcel
• Berntorp Karl
• Di Cairano Stefano

Classifications

• G05B23/0254 Primary

  based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks · CPC title

• G05B13/027

  using neural networks only · CPC title

• G05B13/042

  in which a parameter or coefficient is automatically adjusted to optimise the performance · CPC title

• G05B13/048 Primary

  using a predictor · CPC title

• G05B23/0294

  Optimizing process, e.g. process efficiency, product quality · CPC title