Target-Multiplicity Analysis of CML Processor for Partially-Correlated χ2 Targets

Our scope in this paper is to treat the problem of detecting what is called moderately fluctuating targets when the operating environment is contaminated with a number of outlying targets along with the target under test (mult iple -target situations). The illu mination of this impo rtant class of radar targets by a coherent pulse train, return a train of correlate d pulses with a correlation coefficient in the range 0<ρ<1 (intermediate between SW II & SWI). To achieve this goal, we choo se the OS based type of adaptive detectors owing to its immunity to interfering targets. However, the homogeneous performance of OS technique is always lower than that of the CA scheme. Therefore, it is preferab le to choose the more efficient version, which co mbines the benefits of these two schemes, of the adaptive detectors. This modified version is known as censored mean-level (CM L) in the literature. It imp lements trimmed averaging of a weighted ordered range samp les. Here, the detection performance of the CM L processor is analyzed on the assumption that the radar receiver collects data fro m M successive pulses and the radar system operates in target multip licity environ ments. The primary as well as the secondary interfering targets fluctuates in accordance with χ 2 fluctuation model. SWI and SWII cases represent the situations where the signal is comp letely correlated and completely decorrelated, respectively, fro m pulse to pulse. Exact exp ressions are derived for the detection and false alarm rate perfor mances in nonideal situations. For weak SNR, it is shown that the processor performance imp roves as the correlation coefficient ρ increases and this occurs either in the absence or in the presence of outlying targets. This behavior is rapid ly changed as the SNR beco mes stronger where the p rocessor performance degrades as ρ increases, and the SWII and SWI models emb race all the correlated target cases.