Control system for distributed sensors and actuators

• US 5,365,423 A
• Filed: 01/08/1992
• Issued: 11/15/1994
• Est. Priority Date: 01/08/1992
• Status: Expired due to Fees

• Current Assignee: Rockwell International Corporation
• Original Assignee: Rockwell International Corporation
• Inventors: Chand, Sujeet
• Primary Examiner(s): Envall, Jr., Roy N.
• Assistant Examiner(s): OAKES, BRIAN
• Application Number:
  US07/818,008
• Time in Patent Office: 1,042 Days
• Field of Search: 364/140, 364/141, 364/148, 364/164, 364/165, 364/138, 364/468, 364/149-151, 371/43, 395/2, 395/68
• US Class Current: 700/12
• CPC Class Codes: G05B 19/0423 Input/output
• CPC Class Codes: G05B 19/054 Input/output
• CPC Class Codes: G05B 2219/24075 Predict control element state changes, event changes
• CPC Class Codes: G05B 2219/25235 Associate a sequence function to each control element, event signature
• CPC Class Codes: G05B 2219/2621 Conveyor, transfert line

Abstract

Conventional Boolean Logic Control is augmented to provide enhanced diagnostics, monitoring, and fail safe operation for dynamic systems having distributed discrete-valued sensors and actuators. A decentralized model of a controlled system defines behavior and timing models for both sensors and actuators, termed Control Elements (CEs). Each CE has a first model for transition from state 0 to 1, and a second model for transition from state 1 to 0. Each behavioral model is defined by an Event Signature comprising a sequence of state changes in neighboring CEs. A continuous evaluation of event signatures is performed to compute a probability that a given CE will change state. An Expectation Function is used to check and enforce the correct behavior of a CE. A statistical temporal model predicts delays in the states of a CE as a function of its previous and current delays. The distributed behavior and on-line timing models are used to detect and diagnose incorrect behavior and failures of decentralized sensors and actuators.

Claims

1. A decentralized system for monitoring and controlling a dynamic process, comprising:

a plurality of distributed control elements, each of said control elements having at least two states;

a communication link connecting at least selected ones of said control elements; and

a plurality of microprocessors corresponding to said plurality of control elements, each of said microprocessors dedicated to a corresponding one of said control elements for time-stamping state changes of said one control element;

communicating said time-stamped state changes to selected others of said control elements;

storing a sequence function associated with said one control element, said sequence function comprising a sequence of at least one other control element state change defining an event signature for said one control element to change state;

computing an expectation function for determining a likelihood of said one control element to change state, said expectation function computed from said sequence function; and

evaluating state changes to identify malfunctions of said one control element based on said state change likelihoods determined by said expectation function.

2. The control system of claim 1, wherein each of said microprocessors is further dedicated to said corresponding one of said control elements for estimating a time delay before each state change in said sequence of control element state changes defining said event signature for said one control element.

3. The control system of claim 2, wherein said distributed control elements comprise sensors and actuators.

4. A method of monitoring and controlling a decentralized system having a plurality of distributed control elements, comprising the steps of:

providing each of said control elements with a dedicated microprocessor, each of said control elements having at least two states;

connecting selected ones of said dedicated microprocessors with a communication link;

time-stamping state changes of each of said plurality of control elements;

communicating said time-stamped state changes of each of said control elements to selected others of said dedicated microprocessors connected by said communication link;

providing a plurality of sequence functions, each of said sequence functions associated with a corresponding one of said control elements and comprising a sequence of at least one other control element state change defining an event signature for said one of said control elements;

generating a plurality of expectation functions for determining likelihoods for each of said control elements to change state, each of said expectation functions computed from a corresponding one of said sequence functions; and

evaluating control element state changes to identify state change errors based on said state change likelihoods determined by said expectation functions.

5. The method of claim 4, further comprising the step of estimating a time delay before each state change in said sequence functions defining said event signatures of said control elements.

6. The method of claim 5, further comprising the step of defining each of said event signatures based on a specified sequence of state changes of control elements connected by said communication link.

7. A method of monitoring and controlling a decentralized system having a plurality of distributed control elements including sensors and actuators, comprising the steps of:

providing a plurality of microprocessors, each of said microprocessors connected to at least one of said control elements, each of said control elements having at least two states;

connecting at least selected ones of said microprocessors with a communication link;

monitoring and time-stamping state changes of said plurality of control elements;

communicating said time-stamped state changes of each of said control elements to selected ones of said microprocessors connected by said communication link;

providing a behavioral model for each of said control elements, each of said behavioral models comprising a specified sequence of control element state changes forming an event signature for a specified one of said control elements to change state;

providing a temporal model for each of said control elements, each of said temporal models comprising a time delay estimate for a specified change of state of one of said control elements;

generating a plurality of expectation functions for determining likelihoods of said control elements to change state, each of said expectation functions derived from a corresponding one of said event signatures; and

evaluating control element state changes and identifying state change errors based on said state change likelihoods determined by said expectation functions.

8. The method of claim 7, wherein the step of providing a temporal model for each of said control elements further comprises statistically determining said time delay estimate for each state change of said of control elements.

9. The method of claim 8, further comprising the step of defining each of said event signatures based on said specified sequence of state changes for predetermined ones of said of control elements having said microprocessors connected by said communication link.

10. The method of claim 9, wherein the step of defining said event signatures further comprises the step of defining subsets of the decentralized system for encoding as said event signatures for said control elements.