Abstract :
A new generalized model for atmospheric radio noise is briefly described and justified by comparison with observed probability distributions. A special case of this generalized model is then compared to the observed noise statistics at the very low frequency end of the spectrum. This model is then applied to the detection of known signals in in the presence of noise to determine the optimal receiver structure. The performance of the detector, specified by the upper bound on the probability of error, is assessed and is seen to depend on the signal shape, the time-bandwidth product, and signal-to-noise ratio. The optimal signal to minimize the probability of error is then determined.
