Multilevel crossing rates for automated signal classification
- 24 March 2005
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
Abstract
An investigation was made of multilevel crossing rates as a means of time series analysis of random signals. Pattern recognition techniques based on the Mahalanobis distance were implemented as a means of evaluating the discriminating power of level crossings. Measurement of multilevel crossing rates was found to be an easily implementable means for detection of changes in general frequency content. Level crossing analysis was also shown to be applicable for the study of conductivity measurements of two-phase flow of air and water, where knowledge of the relationship between amplitude and frequency was beneficial in characterizing the process.Keywords
This publication has 6 references indexed in Scilit:
- Machine Recognition of Abnormal Behavior in Nuclear ReactorsIEEE Transactions on Systems, Man, and Cybernetics, 1977
- Results in the Application of Pattern Recognition Methods to Nuclear Reactor Core Component SurveillanceIEEE Transactions on Nuclear Science, 1974
- Level-crossing problems for random processesIEEE Transactions on Information Theory, 1973
- The Expected Number of Zeros of a Stationary Gaussian ProcessThe Annals of Mathematical Statistics, 1965
- Short-Time Frequency Measurement of Narrow-Band Random Signals by Means of a Zero Counting ProcessJournal of Applied Physics, 1955
- Mathematical Analysis of Random NoiseBell System Technical Journal, 1944