Citation

Abstract

We have investigated a command detector in-lock monitoring strategy that uses N estimates of (SNR)'/* each composed of M samples from both data and error channel outputs. The detector recognizes only two states (in-lock and out-of-lock) and indicates state transition when N successive (SNR)*/* estimates violate a threshold. We give the probabilities of indicating in-lock given the detector is out-of-lock and out-of-lock given in-lock as a function of threshold for (N,M) = (1,10), (2,5), (5,2), (10,1). From these probabilities a threshold compatible with design requirements can be determined.

Details

Volume
42-32
Published
April 15, 1976
Pages
70–73
File Size
366.2 KB