Method and apparatus for classifying a signal waveform under test
AssigneeFluke Precision Measurement Ltd.
InventorsPeter Deverson, Floris Broekema
Life of this patent
Estimating the term…
Abstract
A method of classifying a signal waveform under test for appropriate numerical measurement includes acquiring waveform data indicative of the signal waveform under test, analyzing the waveform data, determining a signal waveform type from the analyzed waveform data, and activating an appropriate numerical measurement of the signal waveform under test based upon the determined signal waveform type. There is also provided a test and measurement instrument configured to carry out any of the disclosed methods.
Claims16
A method of classifying a signal waveform under test for appropriate numerical measurement, comprising:
- acquiring waveform data indicative of the signal waveform under test;
- analyzing the waveform data;
- determining a signal waveform type from the analyzed waveform data;
- and
- activating an appropriate numerical measurement of the signal waveform under test based upon the determined signal waveform type,
- wherein determining the signal waveform type from the analyzed waveform data includes:
- determining if the signal is repetitive;
- determining a number of important bins in an amplitude histogram;
- determining a relative importance of bins of the amplitude histogram;
- and
- selecting a type of signal class based upon the determinations, and
- wherein, for the signal waveform to be classified as at least one type of signal class, the method further comprises performing either or both a data activity analysis in which an amount of signal variation in at least a single cycle of the signal waveform under test is compared to a threshold, and/or a matched filter test in which a portion of the signal waveform under test is compared against a template of a known signal type.
A method of classifying a signal waveform under test for appropriate numerical measurement, comprising:
- acquiring waveform data indicative of the signal waveform under test;
- analyzing the waveform data;
- determining a signal waveform type from the analyzed waveform data;
- and
- activating an appropriate numerical measurement of the signal waveform under test based upon the determined signal waveform type,
- wherein acquiring the waveform data comprises storing waveform data values for at least one complete cycle of the signal under test and a value indicative of a next trigger, and
- wherein the value indicative of the next trigger is an approximate repetition rate of the signal waveform under test or a lack of a value indicative of the next trigger is indicative of the signal waveform under test being a non-repetitive signal.
A method of classifying a signal waveform under test for appropriate numerical measurement, comprising:
- acquiring waveform data indicative of the signal waveform under test;
- analyzing the waveform data;
- determining a signal waveform type from the analyzed waveform data;
- and
- activating an appropriate numerical measurement of the signal waveform under test based upon the determined signal waveform type,
- the method further comprising matched filter testing of the signal waveform under test, wherein matched filter testing comprises comparing a portion, optionally at least a single cycle, of the signal waveform under test against a template of a known signal type, and
- wherein the matched filter testing uses a matched filter defined by the equation
- matched
-
-
- filter
- =
- x
- ·
- u
- T
-
- x
-
-
-
- u
-
- ,
- wherein x is a vector of the measured samples, u is the template vector, in the numerator the inner product is used, uT is the transpose of u and a 2-norm function is used in the denominator, wherein a value of 1 indicates a positive similarity, and a value of −1 indicates a negative similarity.
Dependencies are inferred from the claim text.