Reliability Study in Despiking of ADV Data to Open Channel Flow

Abstract

The major problem in collection of data from ADV is generation of spikes due to misidentification of the frequency of Doppler signals, this occurs when the phase difference between outgoing and incoming pulse lies outside the range of -180° to 180°.This shift is caused when the velocity of flow exceeds the pre-set velocity range or when there is contamination from previous pulses reflected from the surface of complex geometrics. For despiking study, different methods are used such as phase space thresholding (PST) method, RC filter (RCF) method, Tukey 53H method (T53H), acceleration thresholding (AT) method, wavelet thresholding (WT) method. Phase space thresholding method uses the concept of three dimensional Poincare maps in which the variables and their derivatives are plotted against each other to determine the spikes. RC filter is a low pass filter used to detect the spikes. Tukey 53H method uses the principle that a median is robust estimator of the mean to generate a smooth time sequence that can be subtracted from original signal. The acceleration threshold method is detection and replacement with two phases one for negative acceleration and other for positive acceleration. WT Method is similar to the previous method except that all the calculations are done in wavelet phase. A review paper is made on the reliability of the most popular methods that are WT, PST and AT method and finally conclusion is prepared showing the most reliable method for ADV despiking