Electric motor controller

• US 20040135534A1
• Filed: 01/14/2003
• Published: 07/15/2004
• Est. Priority Date: 01/14/2003
• Status: Active Grant

• Current Assignee: Christopher P. Cullen
• Original Assignee: Christopher P. Cullen
• Inventors: Cullen, Christopher P.
• Granted Patent: US 7,437,201 B2
• Time in Patent Office: Days
• US Class Current: 318/609
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: H02P 2209/07 Trapezoidal waveform
• CPC Class Codes: H02P 23/14 Estimation or adaptation of motor parameters, e.g. rotor time constant, flux, speed, current or voltage
• CPC Class Codes: H02P 6/16 Circuit arrangements for detecting position

Abstract

A control system for any type of electric motor that automatically learns the characteristics of the motor and computes a motor model for the motor. The control system uses the computed motor model to produce a closed-loop control design that achieves a particular resolution. The control system also uses the motor model to automatically construct efficient motion profiles for a variety of motion commands. The control system may also include an encoder interface device that provides highly accurate motor position information.

Claims

1. A method for adapting a controller for use with a motor, the method comprising:

applying a control signal that causes the motor to change position over a period of time;

recording the control signal over a period of time;

recording position information of the motor over a period of time resulting from the application of the control signal;

estimating a plurality of derivatives of position based on the recorded position information, wherein the plurality of derivatives includes both a first order derivative and a second order derivative;

computing a motor model comprising the step of fitting the plurality of derivatives of position into a motor model equation that relates the control signal to motion of the motor;

and

supplying the motor model to the controller.

2. The method of claim 1 wherein applying a control signal comprises:

applying a modulation signal to a modulator; and

modulating output of a power source in response to the modulation signal to produce a motor control signal; and

wherein recording the control signal over a period of time comprises recording the modulation signal applied to the modulator.

3. The method of claim 1 wherein the controller is coupled to an unregulated power source, and wherein applying a control signal comprises:

receiving one or more measurements of voltage of the power source;

determining a value of a modulation signal that, when supplied to a modulator coupled to the power source, will produce a desired value for the motor control signal, based on the measurements of voltage of the power source; and

applying a modulation signal to a modulator that results in the desired motor control signal;

and wherein recording the control signal over a period of time comprises;

recording the modulation signal applied to the modulator;

recording the measured voltage of the power source; and

then

combining the recording of the modulation signal and the recording the measured voltage of the power source.

4. The method of claim 2 wherein the modulator is configured as a current controller and the motor control signal is a signal in which current varies with time.

5. The method of claim 2 wherein the modulator is configured as a voltage controller and the motor control signal is a signal in which voltage varies with time.

6. The method of claim 1 wherein recording the control signal over a period of time comprises recording the magnitude of the control signal at a series of discrete times.

7. The method of claim 1 wherein the control signal is a modulation signal applied to a modulator, and wherein the method further comprises:

recording current feedback generated by the modulator in response to the modulation signal applied to the modulator; and

fitting the plurality of derivatives of position into an equation that relates the current feedback to motion of the motor.

8. The method of claim 1, wherein recording position information of the motor over a period of time comprises:

sensing a discrete change in motor position; and

storing both a motor position value and an indication of when the change in position occurred.

9. The method of claim 8, wherein recording position information of the motor over a period of time comprises:

recording, at regular intervals, a motor position value and an indication of when a last previous change in position occurred.

10. The method of claim 1, wherein estimating a plurality of derivatives of position comprises:

selecting a set of recorded position and corresponding time data values;

fitting the selected set of recorded position and time data values into a polynomial that represents position as a function of time;

deriving a plurality of derivatives from the polynomial.

11. The method of claim 10, wherein the polynomial comprises at least a second order polynomial.

12. The method of claim 1, wherein the plurality of derivatives further includes a third order derivative.

13. The method of claim 12, wherein estimating a plurality of derivatives of position includes:

selecting a set of recorded position and corresponding time data values; and

fitting the selected set of recorded position and time data values into a polynomial that represents position as a function of time.

14. The method of claim 10, wherein estimating a plurality of derivatives of position further comprises:

prior to fitting the data values, conditioning the time values such that the plurality of selected time values have a mean of approximately zero and a variance substantially equal to one.

15. The method of claim 1, wherein fitting the plurality of derivatives of position into an equation which relates the control signal to the motion of the motor comprises:

computing a least squares fit of the plurality of derivatives of position and corresponding control signal values into an equation that relates the control signal to motion of the motor.

16. The method of claim 15, wherein fitting the plurality of derivatives of position into an equation which relates the control signal to the motion of the motor further comprises:

prior to computing the least-squares fit, conditioning the set of values for each derivative of position such that each set of values has a mean of approximately zero and a variance substantially equal to one.

17. The method of claim 1, wherein estimating the plurality of derivatives of position comprises:

recording a plurality of motor positions and a corresponding time indication associated with each position; and

estimating a velocity based on at least two recorded motor positions and their associated time indications.

18. The method of claim 1, wherein estimating the plurality of derivatives of position, comprises:

estimating acceleration based on at least three recorded motor positions and associated time indications.

19. The method of claim 1, wherein estimating the plurality of derivatives of position comprises:

recording a plurality of motor positions and a corresponding time indication associated with each position;

fitting a describing function to the plurality of motor positions and corresponding times; and

deriving a plurality of derivatives of position from the describing function.

20. The method of claim 19, wherein the describing function is a sum of sinusoids.

21. The method of claim 1 wherein the motor is selected from the group consisting of DC brush, DC brushless, AC brushless, and inductance motors.

22. The method of claim 1 wherein applying the control signal comprises:

transforming a modulation signal into two or more motor phase modulation signals; and

wherein recording the control signal over a period of time comprises recording the modulation signal.

23. The method of claim 1, further comprising:

computing closed-loop control parameters based on the computed motor model.

24. The method of claim 23, wherein computing closed-loop control parameters comprises:

determining a desired resolution value for controller output for each controlled variable;

assigning a first gain of the controller to the desired resolution value;

calculating a frequency value for the gain set to the desired resolution value; and

then,

computing a second gain of the controller.

25. The method of claim 23, wherein computing closed-loop control parameters comprises:

assigning a plurality of proportional gains to a desired resolution value;

calculating a frequency value for each proportional gain set to the desired resolution value;

selecting a smallest frequency value calculated; and

then,

using the selected frequency value, computing remaining gains of the controller.

26. The method of claim 23, wherein computing closed-loop control parameters based on the computed motor model includes:

setting a gain associated with a state variable to achieve a desired resolution for that state variable.

27. The method of claim 2, further comprising:

determining a maximum feedforward signal.

28. The method of claim 27, wherein determining a maximum feedforward signal comprises:

selecting an initial value for the maximum feedforward signal;

performing a motion function using the selected value;

recording the modulation signal used to perform the motion function;

comparing the peak value of the modulation signal used to perform the motion function with a predetermined range; and

if the peak value of modulation signal used to perform the motion function falls outside the predetermined range, then selecting a different value for the maximum feedforward signal and repeating the steps of performing the motion function, recording the modulation signal and comparing the modulation signal.

29. The method of claim 28 wherein the recorded modulation signal is the sum of a feedforward signal and an error correction signal.

30. A method for controlling a motor with a controller, wherein a motor model that relates a motor control signal applied by the controller to the motor to the motion of the motor is predetermined and a maximum feedforward signal has also been predetermined, the method comprising the machine-implemented steps of:

receiving a command to perform a motor function from a user; and

accessing the predetermined motor model and maximum feedforward signal; and

determining a set of motion profile parameters that specify a motor control signal based on the predetermined motor model such that a motor control signal specified by the motion profile parameters does not exceed the maximum feedforward signal; and

supplying the motion profile parameters to a controller to generate the motor control signal specified by the motion profile parameters.

31. The method of claim 30 further comprising:

receiving one or more constraints limiting operation of the motor; and

wherein determining a set of motion profile parameters that specify a motor control signal is also based on the constraints limiting operation of the motor, such that the motor control signal specified by the motion profile parameters satisfies each of the constraints.

32. The method of claim 31 wherein one of the constraints comprises one of the following:

maximum velocity, maximum acceleration, maximum current, maximum RMS current, or maximum RMS acceleration.

33. The method of claim 30 further comprising:

generating a motor control signal from the set of motion profile parameters.

34. The method of claim 30 further comprising:

modifying the set of motion profile parameters to a predetermined format; and

wherein supplying the motion profile parameters to a controller comprises supplying the modified set of motion profile parameters to a controller to generate the motor control signal specified by the motion profile parameters.

35. The method of claim 30 further comprising:

selecting an acceleration pulse form, and wherein the set of motion profile parameters define an acceleration pulse of the selected acceleration pulse form.

36. The method of claim 30 further comprising:

selecting a motion plan algorithm, and wherein determining a set of motion profile parameters is also based on the selected motion plan algorithm.

37. The method of claim 30, wherein the set of motion profile parameters comprise:

a sequence of acceleration pulses, each pulse specified in terms of duration and peak acceleration.

38. The method of claim 37, wherein the sequence of acceleration pulses is any one of the following types of accelerations pulses:

quadratic, harmonic, trapezoidal, piecewise linear, or cycloidal.

39. The method of claim 30, wherein the set of motion profile parameters comprise a set of peak jerk, peak velocity, and peak acceleration values.

40. The method of claim 30, wherein determining a set of motion profile parameters based on a predetermined motor model comprises:

selecting an initial set of motion profile parameters;

computing a peak feedforward signal using the initial set of motion profile parameters;

comparing the computed peak feedforward signal to the predetermined maximum feedforward signal;

if the computed peak feedforward signal is not within a predetermined amount of the maximum feedforward signal, then iteratively selecting other sets of motion profile parameters until a set of motion profile parameters produces a computed peak feedforward signal that is within a predetermined amount of the maximum feedforward signal.

41. The method of claim 30, wherein the command to perform a motion function comprises a command to move to a specified position.

42. The method of claim 30, wherein the command to perform a motion function comprises a command to move with a specified velocity.

43. The method of claim 30, wherein the command to perform a motion function comprises a command to move to a specified position and with a specified velocity.

44. The method of claim 36, wherein the motion plan algorithm comprise any one of the following:

a pulse-cruise-pulse algorithm, a balanced pulse-cruise-pulse algorithm, or a balanced pulse-pulse algorithm.

45. A method for determining a plurality of gains associated with a plurality of variables of a closed-loop controller, the method comprising the computer-implemented steps of:

assigning a gain associated with one of the controller variables to a desired resolution value;

calculating a frequency value for the gain set to the desired resolution value;

computing the remaining gains of the controller;

supplying the computed gains to the closed-loop controller.

46. The method of claim 45, wherein the closed-loop controller comprises a proportional-integral-derivative controller and the assigned gain comprises a proportional gain.

47. The method of claim 46 further comprising:

assigning a plurality of proportional gains of the closed-loop controller to the desired resolution value;

calculating a frequency value for each proportional gain set to the desired resolution value;

selecting the smallest frequency value calculated; and

using the selected frequency value, computing the remaining gains of the controller.

48. The method of claim 46 wherein the closed-loop controller has a predetermined servo sampling rate and computing the remaining gains of the controller comprises:

computing an integral gain;

determining a second resolution value of the controller using the computed integral gain and current servo sampling rate; and

if the second resolution value is greater than the desired resolution value, then adjusting the servo sampling rate to reduce the second resolution value to at or below the desired resolution value.

49. The method of claim 46 wherein computing the remaining gains for the controller comprises:

determining the number of position error readings required as input in a boxcar derivative function in order to achieve a desired resolution.

50. The method of claim 45 wherein the closed-loop controller is a state space controller.

51. A method for designing a closed-loop controller having a state variable and a gain associated the state variable, the method comprising the computer-implemented steps of:

selecting a desired resolution value for the state variable; and

computing a gain associated with the state variable that achieves the desired resolution.

52. The method of claim 51 wherein the closed-loop controller has a plurality of state variables and a gain associated with each state variable, the method further comprising:

selecting a desired resolution value for each state variable; and

computing a gain associated with each state variable that achieves the desired resolution for that state variable.

53. The method of claim 52, wherein computing a gain associated with each state variable that achieves the desired resolution for that state variable comprises:

calculating a frequency value for two or more gains set to the desired resolution value;

selecting the smallest frequency value calculated; and

using the selected frequency value, computing the remaining gains of the controller.

54. A method for calculating a derivative term in a closed-loop controller that receives position error information at each servo processing interval, the method comprising the computer-implemented steps of:

recording three or more position error readings at different servo processing intervals;

computing a derivative estimate by;

summing two or more of the position error readings; and

subtracting an equal number of different position error readings than were used in the summing step; and

then

computing the derivative term based on the derivative estimate.

55. The method of claim 54 wherein the position error readings are recorded at successive servo processing intervals.

56. A method for determining acceleration of a motor based on position information of a motor received from a position measuring device, the method comprising:

receiving position information from the position measuring device;

recording an indication of motor position when a change in the position information is sensed;

recording an indication of the time at which the change in position information occurred; and

calculating acceleration based on at least three position recordings and the corresponding time recordings.

57. The method of claim 56 wherein recording an indication of motor position when a change in the position information is sensed comprises changing the value of a counter based on the change in the position information.

58. The method of claim 57 wherein the counter is incremented when the position information indicates a change in position in a first direction and decremented when the position information indicates a change in position in a second direction.

59. The method of claim 56 calculating acceleration based on at least three position recordings and the corresponding time recordings comprises:

fitting the position/time pair recordings into at least a second order polynomial that relates motor position to time; and

computing the second derivative of the polynomial.

60. A method for determining jerk of a motor based on position information of a motor received from a position measuring device, the method comprising:

receiving position information from the position measuring device;

recording an indication of motor position when a change in the position information is sensed;

recording an indication of the time at which the change in position information occurred; and

calculating jerk based on at least four position recordings and the corresponding time recordings.

61. The method of claim 60 wherein calculating jerk based on at least four position recordings and the corresponding time recordings comprises:

fitting the position/time pair recordings into at least a third order polynomial that relates motor position to time; and

computing the third derivative of the polynomial.

62. A controller for an electric motor having a position-feedback device that provides position feedback information to the controller, the controller comprising:

a motor model computation module configured to compute a motor model equation that relates controller output to motion of the motor; and

a trajectory planning module configured to receive a motion function command and, in response to receiving a motion function command, compute a set of motion profile parameters based on the motor model equation and a predetermined maximum feedforward signal, such that the motion profile parameters describe a motion having a planned peak feedforward signal approximately equal to a predetermined maximum feedforward signal.

63. The controller of claim 62 wherein the planned peak feedforward signal is within 5% of the maximum feedforward signal.

64. The controller of claim 62 further comprising:

a trajectory generator module configured to receive the motion profile parameters and generate a signal to cause the motor to execute the motion function specified in the motion function command.

65. The controller of claim 64 further comprising:

a closed-loop control device configured to calculate position errors during execution of the motion function command and adjust controller output to compensate for calculated errors.

66. The controller of claim 62 further comprising:

a position-measuring interface device configured to receive the position feedback information and record both a change in position and an indication of time at which the change of position occurred.