Improved fading distribution for mobile radio

Probability distribution functions (PDFs) are discussed for characterising narrowband fading processes in mobile radio communications. In particular, the effects of short-term channel variations (fast fading), and longer term channel variations (shadow fading) are discussed in relation to commonly used PDFs. A channel model is proposed to represent the mixture of these processes, and is shown to be well represented by the Suzuki PDF. Difficulties in using the Suzuki PDF to predict system behaviour, in particular bit error rate (BER), are presented, and the use of the generalised gamma PDF is recommended. It is shown that this PDF represents channel behaviour as well as the Suzuki PDF, and that it is much more convenient for BER prediction. The accuracy of BER prediction using this PDF is demonstrated by computer simulation of digital radio transmission over Suzuki fading channels.