Research On Methods For Low Speed Target Detection Of Radar

The modern air threat, in addition to traditional high speed or high mobility targets such as various kinds of bombers, fighters and missiles etc as well as low speed targets such as UAV and gunship etc, have formed new air threat and requested radar detect low speed targets. In consideration of the low speed target generally flies in the low altitude, the clutter should never be neglacted. Generally speaking, the strength of clutter is so big that the radar has to suppress clutter before detecting target. Furthermore, suppression of clutter in frequency-domain processing also suppresses target signal as the result of mixture of low speed targets and clutter in Doppler-domain. It is difficult to improve signal to clutter ratio(SCR), which is the reason of difficulty of detecting low speed target. To solve this problem, the methods of detecting low speed target are studied in this thesis.First, traditional method basing on Kalmus filter and Clutter Map is introduced. At the same time of suppressing clutter Kalmus filter reserves low-frequency components of signal as far as possible, which improves SCR effectively. Comparing with FIR filter, Kalmus filter is more suitable for detecting low speed target. Clutter Map utilizes clutter’s quasi-stationary in time-domain to detect low speed target via so-called time-domain constant false alarm rate(CFAR). Its performance is analyzed, and a program of radar detecting low speed target is given in combination of Kalmus filter. Then several modern signal processing methods, including AR spectral estimation, Capon spectral estimation, super-resolution method and EMD, are studied and analyzed on performance and disadvantage. Comparing with the conventional method based on the Fourier transform, modern methods enable acquisition of higher resolution hopefully. Additionally EMD is a data-adaptive method. Then the method of detecting the hovering helicopter which is a kind of typical low speed target and possesses the main rotor echo of particular characteristic is studied. Through analyzing rotor echo model, this thesis discussed the request for radar to detect the main rotor of helicopter, proposed a time-domain dual threshold detection method and analyzed its performance. Finally, long-time coherent integration methods for rotor echo are introduced and their performance is analyzed.