Method and apparatus for identifying that one of a set of past or historical events best correlated with a current or recent event

• US 4,939,683 A
• Filed: 05/19/1989
• Issued: 07/03/1990
• Est. Priority Date: 05/19/1989
• Status: Expired due to Fees

• Current Assignee: Oh, Seho, Pieter J Van Heerden, Robert J. Marks II
• Original Assignee: Oh, Seho, Pieter J Van Heerden, Robert J. Marks II
• Inventors: Oh, Seho, Marks, Robert J. II, van Heerden, Pieter J.
• Primary Examiner(s): Malzahn, David H.
• Application Number:
  US07/355,202
• Time in Patent Office: 410 Days
• Field of Search: 364/728.03, 364/728.06, 364/724.11, 364/715.11, 382/42
• US Class Current: 708/212
• CPC Class Codes: G06F 17/15 Correlation function computation including computation of convolution operations arithmetic circuits for sum of products per se, e.g. multiply-accumulators G06F7/5443; digital filters, e.g. FIR, IIR, adaptive filters H03H17/00
• CPC Class Codes: G06F 7/02 Comparing digital values G06F7/06, G06F7/22, G06F7/38 take precedence

Abstract

Methods and systems that employ electronic data processing to identify a set of data which can be outputted to a mechanical, electromechanical, electronic, or other device to optimize the performance of that device. The data to be outputted is identified by writing a function which represents a current event or situation as a string of binary digits, finding that one of a series of functions similarly representing past or historical events which is most closely correlated with the current event-representing function, and outputting the string of digits representing the thus identified historical event. The functions representing historical events can be derived from the function representing the current event or from information supplied from a separate source. The system includes delay circuits for generating the historical event-representing strings of digits, circuits for adding the string of digits representing the current situation to the string of digits representing each of the historical events by modulo two addition and counting the number of excess zeros in each resulting sum, a single elimination array of comparators for matching successive pairs of sum-representing strings of digits in enough levels to identify that string of digits with the largest number of excess zeros, and circuitry for: converting the winning string of digits back to the format in which the associated historical event is represented and outputting the string of digits in that format.

Claims

1. A system for generating a set of data which can be inputted to a mechanical, electromechanical, electronic, or other device to control the operation of the device, said system comprising:

a. an array of function comparing means for comparing by modulo two addition a function f(t) representing a current event and presented as a string of binary digits to each of a series of functions D¹ f(t), D² f(t) . . . D^k f(t) representing past or historical events and similarly presented to thereby generate a series of binary digit strings representing the series of sums (1+D¹)f(t), (l+D²)f(t) . . . (l+D^k)f(t);

b. a multiplicity of comparators arrayed in multiple levels with;

the comparators in the first and any intermediate levels arrayed in parallel, the comparators in each level except for the first level having as inputs the outputs from an adjacent pair of comparators in the preceding level, their being a single comparator in the final level, each comparator in the first level having as inputs the outputs from two of said function comparing means which generate successive ones of said series of sums, and each said comparator comprising means for identifying and outputting that one of the two inputted string of binary digits having the greater number of excess zeros where the number of excess zeros in a sum-associated string of digits is that number of zeros which exceeds one-half of the number of digits in the string, whereby the string outputted from the comparator in the final level of the comparator array will be the string of binary digits with the largest number of excess zeros; and

c. means for comparing the string of digits representing the function f(t) to the string of digits with largest number of excess zeros by modulo two addition and for outputting the resulting string of digits from the system.

2. A system as defined in claim 1 which includes means for generating the functions D¹ f(t), D² f(t) . . . D^k f(t) in the form of binary digit strings from the function f(t) and for outputting the thus generated binary digit strings to said function comparing means.

3. A system as defined in claim 2 which has multiple inputs and switching means which can be utilized to connect a selected one of said multiple inputs to said array sum of function comparing means.

4. A system as defined in claim 1 which has an up/down counter between each of the function comparing means in said array and the computer to which that function comparing means outputs the sum generated thereby for ascertaining the number of excess zeros in the string of binary digits representing said sum.

5. A system as defined in claim 1 which has means for enabling said array of function comparing means to accept binary digit strings from which the past event-related functions D¹ g(t), D² g(t) . . . D^k g(t) can be generated from a source independent of the current event-related function f(t).

6. A system as defined in claim 1 in which said function comparing means and said comparators also comprise means for determining which of the past event-related functions D^k f(t) in the more distant past of a current event has the highest degree of correlation with an event in the more recent past of the event or with independently acquired historical events and outputting the string of binary digits representing the thus determined, most highly correlated event.

7. A method of identifying and making available as a string of binary digits that one of a series of historical event-related functions D¹ f(t), D² f(t) . . . D^k f(t) which has the highest degree of correlation with a function f(t) representing a current event or situation, said method comprising the steps of:

a. presenting said historical event-related and current event-related functions as strings of binary digits; and

b. employing an electronic data processing system to;

(i) add the string of binary digits representing the function f(t) to each of the binary digit strings representing a historical event-related function D¹ f(t), D² f(2) . . . D^k f(t) in the modulo two form of addition to thereby produce strings of binary digits representing the resulting sums (1+D¹)f(t), (1+D²)f(t) . . . (1+D^k)f(t),

(ii) determine the excess number of zeros in each string of binary digits representing a sum (1+D¹)f(t), (1+D²)f(t) . . . (1+D^k)f(t) where the number of excess zeros in each sum-representing string of digits is the number of zeros in the string in excess of one-half of the total number of digits in the string;

(iii) compare the number of excess zeros in those binary digit strings representing the sums (1+D¹)f(t)/(1+D²)f(t) . . . (1+D^k-1)f(t)/(1+D^k)f(t) of current event-representing function f(t) and adjacent function pairs D¹ f(t)/D² f(t) . . . D^k-1 f(t)/D^k f(t) and pairs of strings thus determined to have the larger number of excess zeros through the number of levels required to identify that binary digit string with the largest number of excess zeros; and

(iv) output as representative of the past event-related function having the highest degree of correlation with the current event-related function that string of binary digits representing a sum (1+D¹)f(t) . . . (1+D^k)f(t) and having the largest number of excess zeros.

8. A method as defined in claim 7 which includes the step of deriving each of the past event-representing functions D¹ f(t), D² f(t) . . . D^k f(t) from the current event-representing function f(t).

9. A method as defined in claim 8 which includes the step of acquiring the binary digits representing the current event-related function f(t) successively, the step of employing as the string of binary digits representing the first past event-representing function D¹ f(t) the string representing the function f(t) less the last acquired digit in that string, and the step of employing as each succeeding function Df(t) a string of binary digits which is the same as the preceding string less the most recently acquired digit present in the latter.

10. A method as defined in claim 7 which includes the step of deriving the strings of binary digits representing each of the past event-related functions from data acquired from a source independent of that string of digits representing the current event-related function f(t).

11. A method of providing a set of data for optimizing the operation of a mechanical, electromechanical, electronic, or other device which includes the steps of:

a. identifying and making available a string of binary digits representing the sum (1+D^k)f(t) of a current event-related function f(t) and that one of a series of historical event-related functions D¹ f(t), D² f(t) . . . D^k f(t) having the highest degree of correlation with the function f(t) by presenting said historical event-related and current event-related functions as strings of binary digits and employing an electronic data processing system to;

(i) add the string of binary digits representing the function f(t) to each of the binary digit strings representing a historical event-related function D¹ f(t), D² f(2) . . . D^k f(t) in the modulo two form of addition to thereby produce strings of binary digits representing the resulting sums (1+D¹)f(t), (1+D²)f(t) . . . (1+D^k)f(t);

(ii) determine the excess number of zeros in each string of binary digits representing a sum (1+D¹)f(t), (1+D²)f(t) . . . (1+D^k)f(t) where the number of excess zeros in each sum-representing string of digits is the number of zeros in the string in excess of one-half of the total number of digits in the string; and

(iii) compare the number of excess zeros in those binary digit strings representing the sums (1+D¹)f(t)/(1+D²)f(t) . . . (1+D^k-1)f(t)/(1+D^k)f(t) of current event-representing function f(t) and adjacent function pairs D¹ f(t)/D² f(t) . . . D^k-1 f(t)/D^k f(t) and pairs of strings thus determined to have the larger number of excess zeros through the number of levels required to identify that binary digit string (1+D^ko)f(t) with the largest number of excess zeros;

b. adding the strings of digits representing the current event-related function f(t) to the strings of digits representing the aforesaid sum (1+D^ko)f(t) to thereby obtain the string of binary digits representing that historical event-related function having the highest degree of correlation with the function f(t), and

c. outputting to the device the string of binary digits representing the historical event with the highest degree of correlation to the current event and obtained as specified in step b. above.

12. A method of identifying and making available that one of a series of historical event-related functions D^m+1 f(t), D^m+2 f(t) . . . D^k f(t) which has the highest degree of correlation with a function D^m f(t) representing a more recent past of a current event or situation that can be represented by the function f(t), m being smaller than k and said method comprising the steps of:

a. presenting said historical event-related and recent event-related functions as strings of binary digits; and

b. employing an electronic data processing system to;

(i) add the string of binary digits representing the function D^m f(t) to each of the binary digit strings representing a function D^m+1 f(t), D^m+2 f(t) . . . D^k f(t) in the module add two form of addition to thereby produce strings of binary digits representing the resulting sums (1+D^m+1)f(t), (1+D^m+2)f(t) . . . (1+D^k)f(t);

(ii) determine the excess number of zeros in each string of binary digits representing a sum (1+D^m+1)f(t), (1+D^m+2)f(t) . . . (1+D^k)f(t) where the number of excess zeros in each sum representing string of digits is the number of zeros in the string in excess of one-half of the total number of digits in the string;

(iii) compare the number of excess zeros in those binary digit strings representing the sums (1+D^m+1)f(t)/(1+D^m+2)f(t) . . . (1+D^k-1)f(t)/(1+D^k)f(t) of recent event-related function D^m f(t) and adjacent function pairs (D^m+1)f(t)/(D^m+2)f(t) . . . (D^k-1)f(t)/(D^k f(t) and pairs of strings thus determined to have the larger number of excess zeros through the number of levels required to identify that binary digit string with the largest number of zeros; and

(iv) provide as the past event-related function having the highest degree of correlation with the recent event-related function that past event-representing string of binary digits having the largest number of excess zeros.

13. A method of providing a set of data for optimizing the operation of a mechanical, electromechanical, electronic, or other device which includes the steps of:

a. identifying and making available a string of binary digits representing the sum (1+D)f(t) of a recent event-related function D^m f(t) and that one of a series of historical event-related functions D^m+1 f(t), D^m+2 f(t) in which m is smaller than k having the highest degree of correlation with the function D^m f(t) by presenting said historical event-related and current event-related functions as strings of binary digits and employing an electronic data processing system to;

(i) add the string of binary digits representing the function D^m f(t) to each of the binary digit strings representing a function D^m+1 f(t), D^m+2 f(t) . . . D^k f(t) in the modulo two form of addition to thereby produce strings of binary digits representing the resulting sums (1+D^m+1)f(t), (1+D^m+2)f(t) . . . (1+D^k)F(t);

(ii) determine the excess number of zeros in each string of binary digits representing a sum (1+D^m+1)f(t), (1+D^m+2)f(t) . . . (1+D^k)f(t) where the number of excess zeros in each sum representing string of digits is the number of zeros in the string in excess of one-half of the total number of digits in the string; and

(iii) compare the number of excess zeros in those binary digit strings representing the sums (1+D^m+1)f(t)/(1+D^m+2)f(t) . . . (1+D^k-1)f(t)/(1+D^k)f(t) of recent event-related function D^m f(t) and adjacent function pairs D^m+1 f(t)/(D^m+2)f(t) . . . (D^k-)f(t)/D^k f(t) and pairs of strings thus determined to have the larger number of excess zeros through the number of levels required to identify that binary digit string (1+D)f(t) with the largest number of zeros;

b. adding the strings of digits representing the recent event-related function D^m f(t) to the string of digits representing the aforesaid sum (1+D)f(t) having the highest degree of correlation with the recent event-related function D^m f(t) by modulo two addition to thereby obtain the string of binary digits representing that historical event-related function having the highest degree of correlation recent event-related with the function D^m f(t); and

c. outputting to the device the string of binary digits representing the historical event with the highest degree of correlation to the current event and obtained as specified in step b. above.