Copy correlation direction-of-arrival estimation performance

The mean squared error performance of a copy weight correlation (CC) based method of angle estimation proposed by Forsythe and Richmond is studied via application of the method of interval error (MIE). It is demonstrated that at low SNRs the CC method leans heavily on the a priori information embedded in the copy weight, but at high SNRs reverts to the Capon algorithm estimate. This prior information can be quite advantageous when accurate (e.g. it can mitigate the threshold effect altogether), but possibly detrimental if either inaccurate or difficult to match (e.g. due to poorly known array calibration). It is shown here that the error probabilities used in MIE capture the threshold effects induced by mismatch, and correctly predict MSE performance over a wide range of scenarios.