Data processor for executing a fuzzy logic operation and method therefor

• US 5,561,738 A
• Filed: 03/25/1994
• Issued: 10/01/1996
• Est. Priority Date: 03/25/1994
• Status: Expired due to Term

• Current Assignee: Freescale Semiconductor, Inc. (NXP Semiconductors NV)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Pena, David A., Kinerk, Keith E., Quan, Hoang K., Magliocco, Joseph P.
• Primary Examiner(s): Moore, David K.
• Assistant Examiner(s): HAFIZ, TARIQ R
• Application Number:
  US08/217,786
• Time in Patent Office: 921 Days
• Field of Search: 395/3, 395/10-11, 395/51, 395/61, 395/900, 395/76, 395/60, 326/59, 364/746.2, 364/773, 364/800
• US Class Current: 706/4
• CPC Class Codes: G06N 7/04 Physical realisation
• CPC Class Codes: Y10S 706/90 Fuzzy logic

Abstract

A fuzzy inference engine (10) performs fuzzy logic operations with a high degree of accuracy in a minimal amount of time. The fuzzy inference engine (10) includes a fuzzification module (12) which decodes an input signal to access a memory location (18, 20). When accessed, the memory location provides a unary value to a rule evaluation module (26). The rule evaluation module (26) subsequently processes the unary value to indicate a relative strength of a fuzzy inference rule. Because the fuzzy logic operations use unary numbers, rather than more traditional binary numbers, the only time required to perform fuzzy logic operations is equal to only a time required for the signal to propagate through the logic gates (28, 30, 32, 34, 36, and 38) forming rule evaluation module (26).

Claims

1. A method for performing a fuzzy logic operation,

comprising the steps of:

receiving a first input signal;

decoding the first input signal to access a first plurality of memory locations in a first membership function memory when the first input signal is included in the first membership function;

retrieving a first fuzzy input value in unary form from a first one of the first plurality of memory locations, the first fuzzy input value indicating a degree of membership of the first input signal in a first membership function;

performing a first logic function selectively using the first fuzzy input value to provide a first minimum value; and

performing a second logic function selectively using the first minimum value to provide a first output value.

2. The method of claim 1 wherein the first plurality of memory locations correspond to the first membership function.

3. The method of claim 1 further comprising the step of:

concurrently decoding the first input signal to access a second plurality of memory locations in the first membership function memory, the second plurality of memory locations corresponding to a second membership function.

4. The method of claim 3 further comprising the step of:

retrieving a second fuzzy input value in unary form from a first one of the second plurality of memory locations, the second fuzzy input value indicating a degree of membership of the second input signal in the second membership function.

5. The method of claim 4 further comprising the steps of:

performing the first logic function selectively using the first fuzzy input value and a second fuzzy input value to provide the first minimum value; and

performing the second logic function selectively using the first minimum value and the second fuzzy input value to provide the first output value.

6. The method of claim 1 wherein the first fuzzy input value is equal to zero when the input signal is not included within the first membership function.

7. The method of claim 1 wherein the first logic function is an AND function.

8. The method of claim 1 wherein the second logic function is an OR function.

9. The method of claim 1 wherein the first membership function memory is a nonvolatile memory.

10. The method of claim 1 further comprising the steps of:

receiving a second input signal;

decoding the second input signal to access a first plurality of memory locations in a second membership function memory when the second input signal is included in the second membership function;

retrieving a second fuzzy input value in unary form from a first one of the first plurality of memory locations in the second membership function memory, the second fuzzy input value indicating a degree of membership of the second input signal in a second membership function;

performing the first logic function selectively using the first fuzzy input value and second fuzzy input value to provide a second minimum value; and

performing the second logic function selectively using the first minimum value to provide a second output value.

11. A data processor, comprising:

a first address decoder for receiving a first input signal and decoding the first input signal to provide a first decoded signal when the first input signal is in a first membership function;

a first memory having a first plurality of memory locations, the first memory being coupled to the first address decoder for receiving the first decoded signal, the first memory providing a first fuzzy input value in unary form from a first one of the first plurality of memory locations indicated by the first decoded signal;

first logic means for selectively performing a first logic function using the first fuzzy input value, the first logic means being coupled to the first memory for receiving the first fuzzy input value and the first logic means providing a first minimum value; and

second logic means for selectively performing a second logic function using the first minimum value, the second logic means being coupled to the first logic means for receiving the first minimum value and the second logic means providing a first output value.

12. The data processor of claim 11 wherein the first plurality of memory locations corresponds to the first membership function.

13. The data processor of claim 11 wherein the first memory further comprises a second plurality of memory locations, the first memory being coupled to the first address decoder for receiving the first decoded signal, the first memory providing a second fuzzy input value in unary form from a first one of the second plurality of memory locations indicated by the first decoded signal.

14. The data processor of claim 11 wherein the first fuzzy input value indicates a degree of membership of the first input signal in a first membership function.

15. The data processor of claim 11 further comprising:

a second address decoder for receiving a second input signal and decoding the second input signal to provide a second decoded signal when the second input signal is in a second membership function; and

a second memory having a first plurality of memory locations, the second memory being coupled to the second address decoder for receiving the second decoded signal, the second memory providing a second fuzzy input value in unary form from a first one of the plurality of memory locations indicated by the second decoded signal.

16. The data processor of claim 11 wherein the first memory provides a second fuzzy input value in unary form from a second one of the plurality of memory locations indicated by the first decoded signal.

17. The data processor of claim 16 wherein the first logic function performed by the first logic means determines which of the first fuzzy input value and the second fuzzy input value has a lesser value and provides the lesser value as the first minimum value.

18. The data processor of claim 11 wherein the first fuzzy input value is equal to zero when the input signal is not included within the first membership function.

19. The data processor of claim 11 wherein the first logic function is an AND function.

20. The data processor of claim 11 wherein the second logic function is an OR function.

21. The data processor of claim 11 wherein the first memory is a nonvolatile memory.

22. A method for performing a fuzzy logic operation, comprising the steps of:

receiving a first input signal;

decoding the first input signal to access a first plurality of memory locations in a first membership function memory when the first input signal is included in the first membership function memory, the first plurality of memory locations corresponding to a first membership function;

retrieving a first fuzzy input value in unary form from a first one of the first plurality of memory locations, the first fuzzy input value indicating a degree of membership of the first input signal in the first membership function;

concurrently decoding the first input signal to access a second plurality of memory locations in the first membership function memory, the second plurality of memory locations corresponding to a second membership function;

retrieving a second fuzzy input value in unary form from a first one of the second plurality of memory locations, the second fuzzy input value indicating a degree of membership of the second input signal in the second membership function;

performing a first logic function selectively using the first fuzzy input value and a second fuzzy input value to provide a first minimum value; and

performing a second logic function selectively using the first minimum value and the second fuzzy input value to provide a first output value.

23. The method of claim 22 further comprising the steps of:

concurrently receiving a second input signal;

decoding the second input signal to access a first plurality of memory locations in a second membership function memory when the second input signal is included in the second membership function memory;

retrieving a second fuzzy input value in unary form from one of the first plurality of memory locations in the second membership function memory, the second fuzzy input value indicating a degree of membership of the second input signal in a second membership function;

performing the first logic function selectively using the first fuzzy input value and second fuzzy input value to provide a second minimum value; and

performing the second logic function selectively using the first minimum value to provide a second output value.