Neurocomputing control distribution system

• US 6,173,218 B1
• Filed: 10/15/1997
• Issued: 01/09/2001
• Est. Priority Date: 10/15/1997
• Status: Expired due to Term

• Current Assignee: The Boeing Co.
• Original Assignee: The Boeing Co.
• Inventors: Vian, John L.
• Primary Examiner(s): Cuchlinski, Jr., William A.
• Assistant Examiner(s): ARTHUR JEANGLAUDE, GERTRUDE
• Application Number:
  US08/950,966
• Time in Patent Office: 1,182 Days
• Field of Search: 701/1, 701/58, 701/45, 701/46, 701/59, 706/15, 706/23, 706/25, 706/38, 706/39, 706/44, 706/20, 706/26, 706/27, 244/122.AE, 244/122.AB, 244/122.R
• US Class Current: 701/1
• CPC Class Codes: G05B 13/027 using neural networks only
• CPC Class Codes: G05B 2219/33027 Artificial neural network controller
• CPC Class Codes: G05B 2219/33037 Learn parameters of network offline, not while controlling system
• CPC Class Codes: G05B 2219/33043 Extended kalman filter
• CPC Class Codes: G05D 1/0825 using mathematical models

Abstract

A method and system for controlling a propulsion system of a vehicle. The propulsion system includes at least one propulsion effector. The system also includes sensors for sensing vehicle position and motions, a control device for generating control signals, and a generator for generating desired vehicle forces/moments from the sensed vehicle position and motions and the generated control signals based on predefined vehicle compensation and control laws. Also included is a neural network controller for generating propulsion commands for the at least one propulsion effector based on the generated desire forces/moments, wherein said neural network controller was trained based on pregenerated vehicle control distribution data.

Claims

1. A control system for controlling a propulsion system of a vehicle, said propulsion system including at least one propulsion effector, said control system comprising:

at least one sensor for sensing vehicle position and motions;

control means for generating control signals;

a generator for generating desired vehicle forces from the sensed vehicle position and motions and the generated control signals based on predefined vehicle compensation and control laws; and

a neural network controller for generating propulsion commands for the at least one propulsion effector by propagating the generated desired vehicle forces through the neural network controller.

2. A method for controlling a propulsion system of a vehicle, said propulsion system including at least one propulsion effector, said method comprising:

sensing vehicle position and motions;

generating control signals based on control inputs;

generating desired vehicle forces from the sensed vehicle position and motions and the generated control signals based on predefined vehicle compensation and control laws; and

generating propulsion commands for the at least one propulsion effector by propagating the generated desired vehicle forces through a neural network.

3. A method for training a neural network to control a propulsion system of a vehicle, said propulsion system including at least one propulsion effector, said method comprising:

generating sets of desired vehicle forces for possible vehicle command signals, position and motions based on predefined vehicle compensation and control laws;

generating sets of feasible propulsion system forces that correspond to sets of generated desired vehicle forces;

establishing a neural network structure; and

generating nodal weight values for the established neural network structure based on the generated sets of generated desired vehicle forces and the corresponding sets of feasible propulsion system forces.

4. The method of claim 3, wherein said generated sets of feasible propulsion system forces including at least one propulsion effector comprises propulsion effector commands.

5. The method of claim 3, wherein said generating sets of feasible propulsion system forces is based on a linear iterative search algorithm and propulsion system limitations.

6. The method of claim 5, wherein the linear iterative search algorithm proportionally deviates desired forces to generate propulsion commands.

7. The method of claim 5, wherein the linear iterative search algorithm deviates in a prioritized order from the desired forces.

8. The method of claim 3, wherein the established neural network structure is a unified structure.

9. The method of claim 3, wherein the established neural network structure is a decoupled structure.

10. The method of claim 3, wherein the neural network structure comprises neurons with activation functions.

11. The method of claim 10, wherein the activation functions are sigmoidal functions.

12. The method of claim 10, wherein the activation functions are radial basis functions.

13. The method of claim 10, wherein the activation functions are linear functions.

14. The method of claim 3, wherein generating nodal weight values for the established neural network structure further comprises generating nodal weight values based on dynamic derivatives and a node decoupled extended Kalman filter technique.

15. The method of claim 3, wherein generating nodal weight values for the established neural network structure further comprises generating nodal weight values based on a gradient descent technique.

16. The method of claim 3, wherein the neural network structure is implemented as an analog electronic circuit.

17. The method of claim 3, wherein the neural network structure is implemented as a computer program.

18. The method of claim 3, wherein the neural network structure is implemented as an application specific integrated circuit.

19. The method of claim 3, wherein said generating sets of feasible propulsion system forces is based on an optimization algorithm with an objective function and propulsion system limitations.

20. The method of claim 3, wherein said generating sets of feasible propulsion system forces is based on a multiple step reduction algorithm with propulsion system limitations.