Abstract
This paper addresses the problem of small slow target detection in a strong clutter. By exploiting the characteristics of radar cross section fluctuation, the complex Gaussian procedure is utilized to model the matching filter output, introduce the Euclidean distance between the probability density functions (PDF-ED) to represent the resolvability of different hypotheses, and propose an adaptive waveform optimization algorithm that maximizes the PDF-ED between different hypotheses. Compared to the traditional ED between hypotheses, which only takes the mean value of the probability density into consideration, the PDF-ED metric considers the shape of the PDF curve and specifies the essential difference between two hypotheses in terms of information geometry. In this way, the proposed algorithm enhances the target and suppresses the clutter by reducing the side lobes at specified delay-Doppler units, resulting in the improvement of detection performance. The numerical experiments validate the effectiveness of the algorithm.
Highlights
Small slow target detection in a strong clutter is an issue of wide concern, which is very important for a radar system
We adopt the Euclidean distance between the probability density functions (PDF-ED) to scitation.org/journal/adv measure differences between the hypotheses, which greatly simplified the calculation at the expense of a slight acceptable performance penalty for detection compared to Fisher information distance (FID)
The idea of the algorithm we proposed in this paper is that we first construct the fitness function according to the PDF-ED between target-absent and targetpresent hypotheses; we utilize the crossover operation and mutation operation of genetics to adjust the fitness function
Summary
Small slow target detection in a strong clutter is an issue of wide concern, which is very important for a radar system. The target with a small radar cross section (RCS) and low speed tends to be submerged in the strong clutter and cannot be detected effectively. Suppressing side lobes is an effective way to suppress the clutter to improve detection performance, which dates back to the 1950s.4. Side lobes can be reduced from both filter design and waveform design.. Cheng introduced the Fisher information distance (FID) to measure the differences between different hypotheses.. Scitation.org/journal/adv measure differences between the hypotheses, which greatly simplified the calculation at the expense of a slight acceptable performance penalty for detection compared to FID.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call