Comparison of techniques for estimating the frequency selectivity of bandlimited channels

Abstract

A transmission channel used in application such as telecommunications can be modeled as a bandpass filter. Measurement of the frequency selectivity of the channel is important to ensure that the information-bearing signal has minimal distortion and loss of information. A comparison is made for several methods used for estimating the frequency selectivity of the transmission. The methods presented are the correlation method, instantaneous energy and frequency estimation and the cross Wigner-Ville distribution. The theoretical foundations and assumptions are described for each method. In general, all the methods gave similar performance in terms of the frequency selectivity. Due to the shorter analysis duration, both the instantaneous energy and frequency estimation and cross Wigner-Ville distribution is ideal for estimating the frequency selectivity of time-varying channels