On the Problems of Symbol-Spaced Tapped-Delay-Line Models for WSSUS Channels

This paper reviews the pertinence and statistical behavior of symbol-spaced tapped-delay-line (TDL) models which are widely used to model wide-sense stationary uncorrelated scattering (WSSUS) channels. Symbol-spaced TDL models are obtained by sampling the channel impulse response (CIR) at a rate equal to the reciprocal of the symbol duration. They were proposed more than three decades ago within the context of band-limited systems and their applicability seems to be unquestionable. Nonetheless, we show here that these TDL models should be used with care to model WSSUS channels within such a context, because that would violate the channel's uncorrelated scattering (US) condition. We also show that symbol-spaced TDL models suffer from strong limitations in emulating the channel frequency correlation function (FCF). It is shown that this drawback leads to an inaccurate performance evaluation of wireless communication systems sensitive to the FCF. We discuss a simple solution to this problem by doubling the channel's sampling rate. The advantages of this solution will be demonstrated by some numerical examples.