Estimation of error rates and fade distributions on a Rayleigh fading channel with additive white Gaussian noise

Several characteristics of the Rayleigh fading channel are examined. A digital Rayleigh fading simulator is used to generate the (fading) signal envelope from which various statistics are derived. Based on the simulation results, a simple model is proposed in order to estimate the block error rate of a block of N data bits transmitted over the Rayleigh fading channel in the presence of additive white Gaussian noise. This model gives an average estimation error of about 4 % over the range of blocksizes N = 63, 127, 255, 511, 1023, 2047 (bits), average signal-to-noise ratios 70 = 5 to 35 (dB) and fading frequencies f[sub D] = 10 to 90 (Hz) corresponding to vehicle speeds of 8 to 71 MPH at a radio carrier frequency of 850 MHz. A second somewhat more complex model for estimating the block error rate is found to yield a lower average estimation error of 2.4 % over the same set of simulated data. The probability distributions of the fade rate and the fade duration are also examined. Empirical models are derived for the estimation of the probability mass function of the fade rate and the probability density function of the fade duration. These empirical models allow fairly accurate estimates without the need for cosdy and time-consuming simulations. The probability of m-bit errors in an N-bit block is an important parameter in the design of error-correcting codes for use on the mobile radio channel. However, such probabilities are difficult to determine without performing extensive simulation or field trials. An approach to estimate them empirically is proposed.Applied Science, Faculty ofElectrical and Computer Engineering, Department ofGraduat