Research On Sparse Modeling Of Radar Target Signal

The sparse representation can express the original signal with as few basisfunctions as possible in transform domain and extract the most intrinsic characters ofthe signal, so it has been widely used in the signal processing fields such as signalcompression, denoising, super-resolution construction and so on.This article focuses on modeling of radar target signal based on sparserepresentation theory, and mainly consists three parts: the first part studies the sparserepresentation, and discusses redundant dictionary construction of radar signal in orderto match the original echo signal most. Algorithms of matching and basis pursuit areinvestigated and the LASSO algorithm is introduced as a sparse representation method.The second part gets the sparse representation of original echo signal of pulsecompression and phased array radar in order to estimate the distance, velocity and DOAof targets. Computer simulation results indicate that radar target parameter estimationbased on sparse representation is better than traditional method. In the third part, echosignal is sampled at sub-Nyquist sampling rate after sparse representation. Thus we canestimate the target parameters with fewer samples. When estimating DOA of arraysignal, spatial and time domain modeling are adopted, both of which can lower thesampling rate.