Radar target detection method based on Riemannian manifold dimension reduction

Abstract

The invention provides a radar target detection method based on Riemannian manifold dimensionality reduction. The radar target detection method comprises the following steps: acquiring radar echo datareceived by each distance unit; constructing a Riemannian manifold according to the radar echo data; according to the Riemannian manifold, selecting geometric distance measurement, obtaining detection statistics before dimensionality reduction, and then designing a cost function of Riemannian manifold dimensionality reduction; solving a minimization cost function under orthogonal constraints so that a dimension reduction problem can be converted into an optimization problem on a Grassmann manifold; obtaining a Riemannian manifold dimensionality reduction mapping matrix by solving an optimization problem; acting the mapping matrix on Riemannian manifold to realize dimension reduction; and finally, completing the detection of the constant false alarm rate of the matrix on a low-dimensionaldimensionality reduction Riemann manifold with higher discrimination. According to the method, detection performance loss caused by energy leakage of fast Fourier transform is avoided, and meanwhile it is guaranteed that good detection efficiency and detection performance are achieved when the dimension of matrix data is high.

Description

technical field [0001] The invention relates to the field of signal detection, in particular to radar target detection technology, and more specifically to a radar target detection method based on Riemannian manifold dimension reduction. Background technique [0002] Radar target detection is the process of judging the existence of the target of interest by using the information in the radar echo signal. The traditional unit average constant false alarm rate detector based on Doppler processing (M.A.Richards, Fundamentals of Radar Signal Processing, Second Edition, McGraw-Hill, 2014) is to perform fast Fourier transform processing on the echo slow time dimension data , and perform linear filtering or square-law filtering on the processed data, and finally perform unit average constant false alarm rate detection on the filtered data. But in practice, directly performing fast Fourier transform on echo data containing target range-Doppler information will bring lower Doppler r...

AI Technical Summary

Problems solved by technology

However, when the matrix data dimension is high, its detection efficiency and detection performance will be limited.

Images