Delay domain correlation properties of a wideband radio channel

The autocorrelation function of an impulse response can be used for analysing the behaviour of a radio channel, since it describes the statistical properties of the channel. This is commonly used in the time domain, but this paper presents applications of autocorrelation in the delay domain. Four definitions of autocorrelation are proposed and compared, using both simulations and measured data. Simulations show that magnitude correlation gives a measure of the probability that the next incoming path is at a certain distance from the first. Complex correlation gives either correlation between paths with the same fading pattern or correlation between paths with the same Doppler shift. It is shown that the averaging method has substantial influence on the behaviour of the autocorrelation function. Based on analysis of the measured data, it is concluded that complex correlation methods are preferable over those that employ magnitude correlation. The suggestion of a particular method for a certain application, strongly depends on what is actually meant by “being correlated”