Performance analysis for innovations-based detection

The innovations-based detector (IBD) is a powerful hypothesis testing method which uses autoregressive modeling to detector signals in white noise. Previous studies have determined the distribution of the IBD test statistic under the noisy-only and signal plus noise hypotheses. It is shown that the probability of detection is a function of a measure of spectral shape known as the spectral flatness measure. Using the spectral flatness measure, the detection probability can be predicted using the power spectral density of the signal plus noise.<
>