Probability of error performance of the spread-spectrum mobile communications receiver in a non-Rayleigh fading environment

In previous work to evaluate the probability of error performance of the spread-spectrum mobile communications receiver the usual assumptions about the 900 MHz urban mobile channel were made [1], [2], [3], [4], [5]. These "usual assumptions" are that the highly frequency-selective fading has Rayleigh statistics with a slowly-changing, non-frequency-selective median, the median has log-normal characteristics [6]. This paper studies the effects of the fading distribution on the performance of the Spread-Spectrum receiver. The Spread-Spectrum systems uses transmitter power control to counter the effects of shadow fading in the mobile to base station channel. This study is thus confined to the effects of deviations from the Rayleigh fading assumption. The effects of the distribution of the median are not considered at this time. The design of the signal set used in the spread-spectrum system incorporates inherent diversity into the operation of the receiver. Loose bounds on the probability of error can be set by considering the cases of flat fading over the entire spread-spectrum signal, and independent fading of each component of the spread-spect rural signal. The results of Monte-Carlo computer simulations show that fading distributions that are the result of dominant specular reflections cause deleterious probability of error performance. The result of Jao and Elbaum [7] for single pulse detection in non-Rayteigh fading channels provides the motivation and the genesis for the work presented here.