Circular Correlation Technique for Filtering Randomly Noisy Data

摘要

This paper introduces a technique of smoothing high frequency random fluctuations from noisy data. An optimum lowpass filter is developed by using the circular serial correlation statistic to minimize the random noise content in serial data. The procedure is applicable for statistically mild data. All the basic aspects of the new filter technique are discussed and are applied to both simulated noisy data and actual measured data of unknown noise reduction, but it also appears to provide a means to reconstruct undersampled data.