test on the absolute in-scan
position discrepancy between
the drop-dead and the candidate, in units of the standard deviation
of the discrepancy random variable. Assuming independent errors,
this is the square root of the sum of the in-scan error variances
of the two detections. This test was used primarily for its computational
speed and for the fact that it permitted one to set the threshold
by selecting the fraction of all true events which one was willing
to sacrifice in order to deter false events
(Table V.D.1).
When the coarse in-scan window was exhausted the search for candidates was terminated. At that point, the number of candidates which passed the fine position test with the drop-dead determined the next step. If only one was found the double-detection mode was processed. If two were found, the triple-detection mode was examined. If more than two candidates satisfied the position agreement requirement, then confusion processing was invoked. Each of these three cases is discussed below.
If no acceptable candidates were found or if none remained after the additional testing discussed below, then the drop-dead detection was rejected. This means that it was considered either non-seconds-confirmed (NSC) if all of the compatible redundant detectors were operational, or non-seconds-confirmed due to a failed detector (NSCF) if the alibi of a dead or degraded detector was applicable.