System and method for self-tuning feedback control of a system

• US 6,631,299 B1
• Filed: 12/15/1999
• Issued: 10/07/2003
• Est. Priority Date: 12/22/1998
• Status: Expired due to Term

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Brooks, Clifton E., Hilbun, Stephanie L., Patel, Nital S., Jenkins, Steven T.
• Primary Examiner(s): Follansbee, John
• Assistant Examiner(s): GAIN, EDWARD F
• Application Number:
  US09/464,658
• Time in Patent Office: 1,392 Days
• Field of Search: 700/37, 700/32, 700/28, 700/121
• US Class Current: 700/37
• CPC Class Codes: G05B 13/042 in which a parameter or coefficient is automatically adjusted to optimise the performance

Abstract

A tuned run-to-run controlled system is disclosed that provides tuned run-to-run control of a system. The system includes a controlled system coupled to a tuned run-to-run controller, which contains a feedback controller coupled to a tuner. Tuned run-to-run controller determines a feedback command based on a nominal gain, a maximum gain, a process error, and a tuning gain.

Claims

1. A method for controlling a system, comprising the steps of:

providing a system to be controlled;

determining and storing a set of system parameters;

determining a sequence of error measurements from said system;

determining a stable sequence of tuning gains based upon the sequence of error measurements and the stored system parameters;

determining a feedback offset for said system using the stable sequence of tuning gains; and

providing a system controller for controlling said system responsive to a process measurement from said system and said feedback offset.

2. The method of claim 1 further including the step of generating a control signal responsive to said feedback offset.

3. A method for controlling a system, comprising:

determining a nominal gain of a system;

determining a maximum gain of a system;

determining a sequence of error measurements of the system;

determining a stable sequence of tuning gains based upon the sequence of error measurements, the nominal gain, and the maximum gain; and

tuning the system using the stable sequence of tuning gains;

wherein the stable sequence of tuning gains comprises;

a first stable sequence of tuning gains in a first mode of operation to increase responsiveness to error induced by an unknown disturbance; and

a second stable sequence of tuning gains in a second mode of operation to decrease sensitivity to error induced by an unknown noise.

4. The method of claim 3, wherein each selected error measurement from the sequence of error measurements comprises an error induced by an unknown noise and an error induced by an unknown disturbance.

5. The method of claim 3, further comprising:

in a first mode of operation, determining a stable sequence of tuning gains by selectively weighting a high frequency component of the sequence of error measurements and a low frequency component of the sequence of error measurements; and

in a second mode of operation, determining a stable sequence of tuning gains by disregarding the high frequency component of the sequence of error measurements.

6. The method of claim 5, wherein:

the high frequency component comprises a mean square error term; and

the low frequency component comprises a square of the mean error term.

7. The method of claim 6, wherein the stable sequence of tuning gains biases the mean square error to be minimal at the nominal gain.

8. The method of claim 3, wherein the mean square error of the sequence of error measurements increases asymptotically as a function of the difference between the nominal gain and an actual gain of the system.

9. The method of claim 3, wherein the stable sequence of tuning gains controls a rate of response of the system to the sequence of error measurements.

10. The method of claim 3, wherein:

the stable sequence of tuning gains is driven toward an optimal tuning gain; and

the optimal tuning gain provides an optimal tradeoff between maximizing the responsiveness to a low frequency component of the sequence of error measurements and minimizing the sensitivity to a high frequency component of the sequence of error measurements.

11. A method for controlling a system, comprising:

determining a nominal gain of a system;

determining a maximum gain of a system;

determining a sequence of error measurements of the system;

determining a stable sequence of tuning gains based upon the sequence of error measurements, the nominal gain, and the maximum gain; and

tuning the system using the stable sequence of tuning gains;

further comprising the step of filtering a first error message from the sequence of error measurements based on a comparison between the first error measurement and a previous error measurement from the sequence of error measurements.

12. A self-tuning system, comprising:

a system controller for generating a measurement from a process and controlling said process based on a command;

a metrology device for measuring a result of said process and generating an error signal based on the result measured by said metrology device;

a tuner responsive to said error signal for determining a stable tuning gain based upon the error, a nominal gain for the process, and a maximum gain for the process; and

a feedback controller for determining a feedback target based upon the error signal and said stable tuning gain for generating the command responsive to the measurement and the determined feedback target.

13. A self-tuning system, comprising:

a system controller for generating a measurement and performing a process based on a command;

means to generate an unknown noise and an unknown disturbance;

a metrology device for measuring a result of the process and generating an error signal based on the result measured by the metrology device, said error signal induced by said unknown noise and by said unknown disturbance;

a tuner responsive to said error signal for determining a stable tuning gain based upon the error signal, a nominal gain for the process, and a maximum gain for the process; and

a feedback controller for determining a feedback target based upon the error signal and the stable tuning gain for generating the command based on the measurement and the feedback target, wherein the error signal comprises an error induced by said unknown noise and said error induced by an unknown disturbance.

14. The self-tuning system of claim 13, wherein the stable tuning gain controls a rate of response of the feedback controller.

15. The self-tuning system of claim 13, wherein the stable tuning gain prevents a resulting gain of the process from exceeding the maximum gain.

16. The self-tuning system of claim 13, wherein the tuner is further operable:

in a first mode of operation, to determine a stable tuning gain by selectively weighting a high frequency component of the error and a low frequency component of the error; and

in a second mode of operation, to determine a stable tuning gain by disregarding the high frequency component of the error.

17. The self-tuning system of claim 13, wherein the tuner is further operable to increase the stable tuning gain in response to an increase of the low frequency component of the error.

18. A self-tuning photolithography system, comprising:

a stepper for aligning a target of a photolithography overlay based on an alignment command;

a sensor for generating an alignment measurement signal;

means to generate unknown noise and unknown disturbance signals;

a metrology device to detect an error of the stepper and provide an error signal which comprises a high frequency component induced by said unknown noise and a low frequency component induced by said unknown disturbance;

a tuner responsive to said error signal to determine a stable tuning gain based upon the error, a nominal gain for the stepper, and a maximum gain for the stepper;

a feedback controller to determine a feedback offset responsive to the error signal and the stable tuning gain; and

a stepper alignment controller responsive to said feedback offset and said measurement signal for generating the alignment command.

19. The photolithography system of claim 18, wherein the tuner is further operable:

in a first mode of operation, to determine a stable tuning gain by selectively weighting the high frequency component and the low frequency component; and

in a second mode of operation, to determine a stable tuning gain by disregarding the high frequency component.

20. The photolithography system of claim 18, wherein the stable tuning gain controls a rate of response of the feedback controller.