Probabilistic control of an agricultural machine

• US 9,807,932 B2
• Filed: 10/02/2015
• Issued: 11/07/2017
• Est. Priority Date: 10/02/2015
• Status: Active Grant

• Current Assignee: Deere & Company
• Original Assignee: Deere & Company
• Inventors: French, Jr., William D., Bruns, Aaron J., Kis, Janos
• Primary Examiner(s): Black, Thomas G
• Assistant Examiner(s): HUYNH, LUAT T
• Application Number:
  US14/874,223
• Publication Number:
  US 20170094901A1
• Time in Patent Office: 767 Days
• Field of Search: 701 50, 460 1, 460 4- 6
• CPC Class Codes: A01D 41/127 Control or measuring arrangements specially adapted for combines
• CPC Class Codes: A01D 41/1273 for measuring grain loss
• CPC Class Codes: A01F 12/18 Threshing devices
• CPC Class Codes: A01F 12/444 Fanning means
• CPC Class Codes: A01F 12/446 Sieving means
• CPC Class Codes: A01F 12/46 Mechanical grain conveyors
• CPC Class Codes: G06Q 10/0639 Performance analysis of employees; Performance analysis of enterprise or organisation operations
• CPC Class Codes: G06Q 50/02 Agriculture; Fishing; Forestry; Mining

Abstract

A set of sensor inputs are received in an agricultural machine. The sensor inputs are indicative of sensed or measured variables. A probabilistic control system probabilistically infers values for another set of variables and generates a set of control signals based on the probabilistically inferred values.

Claims

1. An agricultural mobile machine, comprising:

a plurality of different sensors, each sensing a different variable and generating a corresponding sensor signal indicative of the sensed variable;

a plurality of different controllable subsystems;

a probabilistic controller configured to receive the sensor signals and generate a set of estimated variable values for unobserved variables and generate a plurality of different subsystem control signals to control the plurality of different subsystems, based on the sensor signals and the estimated variable values for the unobserved variables, wherein the probabilistic controller comprises a probabilistically inferred value generator that generates the estimated variable values for the unobserved variables;

a control signal generator that generates the plurality of different subsystem control signals; and

a material processing system that processes material in a harvesting operation and

wherein the probabilistically inferred value generator comprises;

a material stream volume estimator component that generates estimated values indicative of material volumes at different places in in the material processing system during the harvesting operation; and

a crop behavior characteristic estimator component that generates estimated values indicative of processing characteristics of a crop being harvested during the harvesting operation.

2. The agricultural mobile machine of claim 1 wherein the agricultural mobile machine comprises a combine.

3. The agricultural mobile machine of claim 2 wherein the control signal generator comprises:

a cost function analysis component that generates the plurality of different subsystem control signals based on a cost function and the estimated values estimated by the material stream volume estimator component and the crop behavior characteristic estimator component.

4. The agricultural mobile machine of claim 2 wherein one of the plurality of subsystems comprises a threshing subsystem that includes a concave and a rotor, and wherein one of the plurality of subsystem control signals comprises:

a concave clearance signal that controls concave clearance in the threshing subsystem.

5. The agricultural mobile machine of claim 4 wherein another of the plurality of subsystem control signals comprises:

a rotor speed control signal that controls rotor speed in the threshing subsystem.

6. The agricultural mobile machine of claim 2 wherein one of the plurality of subsystems comprises a cleaning subsystem that includes a chaffer, a sieve and a fan, and wherein one of the plurality of subsystem control signals comprises:

a chaffer opening control signal that controls a chaffer opening in the cleaning subsystem.

7. The agricultural mobile machine of claim 6 wherein another of the plurality of subsystem control signals comprises:

a sieve opening control signal that controls a sieve opening in the cleaning subsystem.

8. The agricultural mobile machine of claim 7 wherein another of the plurality of subsystem control signals comprises:

a fan speed control signal that controls a speed of the fan in the cleaning subsystem.

9. The agricultural mobile machine of claim 2 wherein one of the plurality of subsystems comprises a machine propulsion subsystem that drives movement of the combine over the ground and wherein one of the plurality of subsystem control signals comprises:

a ground speed control signal that controls the propulsion subsystem.

10. The agricultural mobile machine of claim 2 wherein one of the plurality of subsystems comprises a front end equipment subsystem with a header motor that controls header height of a header and wherein one of the plurality of subsystem control signals comprises:

a header height control signal that controls the header motor in the front end equipment subsystem.

11. The agricultural mobile machine of claim 2 wherein the probabilistic controller comprises:

a Bayesian network controller.

12. The agricultural mobile machine of claim 2 wherein the probabilistic controller models the unobserved variables using a hidden Markov model.

13. A combine, comprising:

a plurality of different sensors, each sensing a different variable and generating a corresponding sensor signal indicative of the sensed variable;

a plurality of different controllable subsystems;

a material processing system that processes material in a harvesting operation; and

a probabilistic controller configured to receive the sensor signals and generate a set of estimated variable values for unobserved variables and generate a plurality of different subsystem control signals to control the plurality of different subsystems, based on the sensor signals and the estimated variable values for the unobserved variables, wherein the probabilistic controller comprises a probabilistically inferred value generator that generates the estimated variable values for the unobserved variables; and

wherein the probabilistic controller comprises;

a material stream volume estimator component that generates estimated values indicative of material volumes at different places in in the material processing system during the harvesting operation;

a crop behavior characteristic estimator component that generates estimated values indicative of processing characteristics of a crop being harvested during the harvesting operation; and

a control signal generator that generates the plurality of different subsystem control signals, based on the sensor signals and the estimated values.

14. A method of controlling a combine, comprising:

sensing a plurality of different variables;

generating a corresponding sensor signal indicative of each of the sensed variables;

probabilistically generating, using a probabilistic controller, a set of estimated variable values for unobserved variables based on the sensor signals, wherein the probabilistic controller comprises a probabilistically inferred value generator that generates the estimated variable values for the unobserved variables, wherein the combine includes a material processing system that processes material in a harvesting operation, and wherein probabilistically generating the set of estimated variable values comprises;

generating estimated values indicative of material volumes at different places in in the material processing system during the harvesting operation; and

generating estimated values indicative of processing characteristics of a crop being harvested during the harvesting operation; and

generating a plurality of different subsystem control signals to control a plurality of different subsystems, based on the sensor signals and the estimated variable values for the unobserved variables.

15. The method of claim 14 wherein probabilistically generating the set of estimated values comprises using a Bayesian network to generate the set of estimated values, using values represented by the sensor signals as evidence variables.