Adaptive detection in stationary and nonstationary noise environments
Abstract
This report describes the statistical performance of several radarbased adaptive detection schemes in both stationary and nonstationary noise and interference environments. The detectors under study must be able to correctly determine the presence of a target in a range gate with a high degree of probability given that the probability of misclassification is a fixed small value. The hostile noise environment is assumed to consist of possibly timevarying, spatially correlated interference along with Gaussian background noise. In a typical radar environment, the mean value of the returned radar signal and the noise covariance matrix are unknown parameters; therefore, generalized likelihood ratio test procedures were used to develop decision rules that meet the NeymanPearson criterion. Three major cases of interest were examined. First, the singlepulse test developed by Kelly is reviewed. The multiplepulse return test case is extremely complicated and was divided into distinct detector forms: noncoherent and coherent. The performance of each detector is a function of the signaltonoise ratio, the number of radar pulse returns used in the decision rule, and the quality of the covariance estimate.
 Publication:

NASA STI/Recon Technical Report N
 Pub Date:
 February 1994
 Bibcode:
 1994STIN...9521099M
 Keywords:

 Background Noise;
 Detection;
 Noise;
 Radar;
 Radar Targets;
 Random Noise;
 Statistical Analysis;
 Noise Measurement;
 Signal To Noise Ratios;
 Communications and Radar