Multiple scale signal processing and control system

• US 6,041,172 A
• Filed: 11/26/1997
• Issued: 03/21/2000
• Est. Priority Date: 11/26/1997
• Status: Expired due to Term

• Current Assignee: Tokyo Electron Limited
• Original Assignee: Voyan Technology Incorporated (HP Inc.)
• Inventors: Shah, Sunil C., Erickson, Mark A., Pandey, Pradeep
• Primary Examiner(s): Teska, Kevin J.
• Assistant Examiner(s): Broda, Samuel
• Application Number:
  US08/980,041
• Time in Patent Office: 846 Days
• Field of Search: 364/578, 392/416, 395/500.23, 395/500.28, 700/29, 700/32
• US Class Current: 703/6
• CPC Class Codes: G05B 13/04 involving the use of models or simulators
• CPC Class Codes: G05B 13/042 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G05B 17/02 electric

Abstract

A method for processing signals and controlling a physical system in which measurements are obtained at different time scales and/or different space scales. Signals generated from the physical system are processed by first creating a first dynamic model at a first time/space scale that consists of a first set of parameters, a first set of states and a first set of inputs. A second dynamic model at a second time/space scale is also created and consists of a second set of parameters, a second set of states and a second set of inputs. At least one of the first set of parameters in the first dynamic model are computed from the second set of states in the second dynamic model. A second estimator is then created to produce estimates of the second set of states using the second dynamic model, measurements from the first physical system at the second time scale. A first estimator is also created to produce estimates of the first set of states using the first dynamic model, measurements from the first physical system at the first time scale, and the estimates from the second estimator.

Claims

1. A method of processing signals generated from a first physical system comprising the steps of:

a) creating a first dynamic model of the first physical system, the first dynamic model being created at a first time scale that consists of a first set of parameters, a first set of states and a first set of inputs;

b) creating a second dynamic model of the first physical system, the second dynamic model being created at a second time scale that consists of a second set of parameters, a second set of states and a second set of inputs, at least one of the first set of parameters in the first dynamic model being computed from the second set of states in the second dynamic model;

c) creating a second estimator to produce estimates of the second set of states using the second dynamic model, measurements from the first physical system at the second time scale; and

d) creating a first estimator to produce estimates of the first set of states using the first dynamic model, measurements from the first physical system at the first time scale, and the estimates from the second estimator.

2. The method of claim 1 wherein the step of creating the second estimator is performed by further using measurements from the first physical system at the first time scale.

3. The method of claim 1 wherein the first set of parameters comprises a first subset of parameters that are constant and a second subset of parameters that vary with time, the second subset of parameters computed from the second set of states in the second model.

4. The method of claim 1 wherein the first dynamic model is created using diagnostic sensors.

5. The method of claim 1, wherein the step of creating the first dynamic model comprises the steps of:

creating a set of test signals that exercise high bandwidth and low bandwidth dynamics of the first physical system;

applying the test signals to the first physical system while capturing a synchronous response of the first physical system to the test signals, the first dynamic model correlating the test signals and synchronous response of the first physical system.

6. The method of claim 1 wherein at least one of the estimators is based on linearized models at multiple operating points.

7. The method of claim 1 wherein at least one of the estimators is based on a continuous-time extended Kalman filter (CTEKF) for state estimation.

8. The method of claim 1 wherein at least one of the estimators is based on a continuous-time extended Kalman filter (CTEKF) for state estimation and the diagnostic measurements.

9. The method of claim 1 wherein at least one of the estimators is based on a discrete-time extended Kalman filter (DTEKF) for state estimation.

10. The method of claim 1 further comprising the step of processing signals generated from a second physical system which comprises the steps of

a) applying the first dynamic model to the second physical system to create a third dynamic model at the first time scale that represents the behavior of the second physical system, the third dynamic model consisting of a third set of parameters, a first set of states and a first set of inputs;

b) creating a fourth dynamic model at the second time scale that consists of a fourth set of parameters, a fourth set of states and a fourth set of inputs, at least one of the third set of parameters in the third dynamic model being computed from the fourth set of states in the fourth dynamic model;

c) creating a fourth estimator to produce estimates of the fourth set of states using the fourth dynamic model and measurements from the second physical system at the second time scale; and

d) creating a third estimator to produce estimates of the third set of states using the third dynamic model, measurements from the second physical system at the first time scale, and the estimates from the fourth estimator; and

d) creating a second controller that operates at the first time scale based on the estimates from the third and fourth estimators, the controller optimizing the performance of the second physical system.

11. The method of claim 10 wherein the step of creating the fourth estimator is performed by further using measurements from the second physical system at the first time scale.

12. A method for controlling a second physical system in a family of physical systems that includes a first physical system and the second physical system, the method comprising the steps of:

a) creating a first dynamic model at a first time scale that indicates the behavior of the first physical system, wherein the first dynamic model consists of a first set of states, a first set of inputs, and a first set of parameters, the first set of parameters comprising a first subset of parameters and a second subset of parameters, the first subset of parameters being common to the first and second physical systems;

b) creating a second dynamic model at a second time scale that indicates the behavior of the first physical system at the second time scale, wherein the second dynamic model consists of a second set of parameters, a second set of states and a second set of inputs, the second subset of parameters in the first dynamic model being computed from the second set of states in the second model;

c) creating a second estimator to produce estimates of the second set of states using the second dynamic model and measurements from the second physical system at the second time scale;

d) creating a first estimator to produce estimates of the first set of states using the first dynamic model, measurements from the first physical system at the first time scale, and the estimates from the second estimator;

e) creating a controller that operates at the first time scale based on the estimates from the first and second estimators, the controller optimizing the performance of the second physical system;

f) applying the first dynamic model to the second physical system to create a third dynamic model at the first time scale that represents the behavior of the second physical system, the third dynamic model consisting of a third set of states, a third set of inputs, and a third set of parameters, the third set of parameters comprising the first subset of parameters and a third subset of parameters;

g) creating a fourth dynamic model at the second time scale that consists of a fourth set of parameters, a fourth set of states and a fourth set of inputs, the third subset of parameters in the third dynamic model being computed from the fourth set of states in the fourth dynamic model;

h) creating a fourth estimator to produce estimates of the fourth set of states using the fourth dynamic model, measurements from the second physical system at the second time scale;

i) creating a third estimator to produce estimates of the third set of states using the third dynamic model, measurements from the second physical system at the first time scale, and the estimates from the fourth estimator; and

j) creating a second controller that operates at the first time scale based on the estimates from the third and fourth estimators, the controller optimizing the performance of the second physical system.

13. The method of claim 12 wherein the step of creating the second estimator is performed by further using measurements from the first physical system at the first time scale.

14. The method of claim 12 wherein the step of creating the fourth estimator is performed by further using measurements from the second physical system at the first time scale.

15. The method of claim 12, wherein the step of creating the first dynamic model comprises the steps of:

creating a set of test signals that exercise high bandwidth and low bandwidth dynamics of the first physical system;

applying the test signals to the first physical system while capturing a synchronous response of the first physical system to the test signals, the first dynamic model correlating the test signals and synchronous response of the first physical system.

16. The method of claim 15, wherein the step of capturing a synchronous response of the first physical system comprises the steps of capturing a series of measurements through a set of diagnostic sensors in the first physical system.

17. The method of claim 15, wherein the step of capturing a synchronous response of the first physical system comprises the steps of capturing a series of measurements through a set of production sensors and a set of diagnostic sensors in the first physical system.

18. The method of claim 12, wherein each first and second physical system comprises an evaporation chamber and a deposition target.

19. The method of claim 12 wherein at least one of the estimators is based on linearized models at multiple operating points.

20. The method of claim 12 wherein at least one of the estimators is based on a continuous-time extended Kalman filter (CTEKF) for state estimation.

21. The method of claim 12 wherein at least one of the estimators is based on a continuous-time extended Kalman filter (CTEKF) for state estimation and the diagnostic measurements.

22. The method of claim 12 wherein at least one of the estimators is based on a discrete-time extended Kalman filter (DTEKF) for state estimation.

23. A method of processing signals generated from a physical system comprising the steps of:

a) creating a first dynamic model of the physical system, the first dynamic model being created in a first space scale that consists of a first set of parameters, a first set of states and a first set of inputs;

b) creating a second dynamic model of the physical system, the second dynamic model being created in a second space scale that consists of a second set of parameters, a second set of states and a second set of inputs, at least one of the first set of parameters in the first dynamic model being computed from the second set of states in the second dynamic model;

c) creating a second estimator to produce estimates of the second set of states using the second dynamic model, measurements from the physical system in the second space scale; and

d) creating a first estimator to produce estimates of the first set of states using the first dynamic model, measurements from the physical system in the first space scale, and the estimates from the second estimator;

e) creating a controller that operates in the first space scale based on the estimates from the first and second estimators, the controller optimizing the behavior of the physical system.

24. The method of claim 23 wherein the first dynamic model is created using diagnostic sensors.

25. The method of claim 23 wherein at least one of the estimators is based on linearized models at multiple operating points.

26. The method of claim 23 wherein at least one of the estimators is based on a continuous-time extended Kalman filter (CTEKF) for state estimation.

27. The method of claim 23 wherein at least one of the estimators is based on a continuous-time extended Kalman filter (CTEKF) for state estimation and the diagnostic measurements.

28. The method of claim 23 wherein at least one of the estimators is based on a discrete-time extended Kalman filter (DTEKF) for state estimation.