adaptive signal processing
Perspective radar must be ready to work in conditions of multiple noises, which present additive mixture of simultaneously acting, having a priori unknown parameters different kind of noises with natural and artificial origin (passive clutter from clouds (PC), local subjects (LS), thrown dipoles, active counteraction (AC) – persistent noise and unsynchronized noise (UN)). For detection of small size radar objects in such situations it’s need to use coherent patches, consisted from enough large number of impulses (8 and more), preferably with stagger of pulse repetition period, processing. Purpose of this work is to research of different construction versions and efficiency analyses of adaptive to unknown multiple noises parameters patches of signals with period repetition stagger processing. Herewith ways to simplify of processing realization in part of adaptive whitening filter (AWF) and signal accumulator (coherent – CA or incoherent – IA) and also possibilities of multi-channel on types and noise parameters processing systems are considered.
Versions of construction AWF-CA and AWF-IA for coherent patch with pulse repetition period stagger in conditions of multiple noises have researched. It’s shown that not requiring multi-channel CA realization processing AWF-IA yields AWF-CA only 0.7-0.9dB with N=8. Herewith to get loss of ~0.5 – 1dB it’s need size of averaging window , consisted 64 – 128 elements. It’s possible to simplify of matrix AWF realization with lattice filter structure and censor of signals, which have got in the averaging window. Additional losses herewith do not exceed 0.5dB, if 64. For type of noises PC+UN the example of multi-channel system, based on simpler adaptive to only value of noise rejecter filter using construction is considered. Loss of these systems to AWF-IA consists up to 6dB because of filtered signals accumulation absence. Together with that loss from finite size of averaging window and errors of processing channel choice have not exceeded ~1dB already with =32.