Research On Heterogeneous Clutter Suppression Method Of Airborne Early Warning Radar

Airborne early warning radar receives scattering clutter from ground and sea surface when operating at look-down mode.Due to the high-speed motion of the platform,the clutter spreads seriously in the Doppler frequency domain,and the moving target may be buried in the clutter.The space-time adaptive processing(STAP)technology can distinguish the clutter and target in spatial-temporal domain,and improve the clutter suppression capability.Statistical STAP algorithms require independent and identically distributed(i.i.d.)samples to estimate the covariance matrix.In the heterogeneous clutter environment which is caused by complex radar working environment,non-sidelooking array and many other factors,it's difficult to get enough i.i.d.samples and the performance will be obviously degraded.Knowledge-aided signal processing is an important heterogeneous clutter suppression method,but current studies didn't fully consider target steering vector mismatch and clutter model mismatch which are caused by array amplitude and phase error and the off-grid of sparse recovery STAP(SR-STAP).So it is needed to research on signal processing method of practical value.This thesis mainly studies on heterogeneous clutter suppression methods for airborne early warning radar.Aiming at the problem that adaptive algorithms hardly maintain the target gain and the signal-to-noise ratio is reduced due to amplitude and phase error of the array,two error estimation methods are studied.The first one builds cost function which integrates error estimation into the filtering process of knowledge-aided clutter suppression,which can obtain estimated error and filtered output of clutter suppression.The second method iteratively estimates error with temporal reduced-dimension processing,which can reduce the calculation load and be used as a preprocessing of traditional STAP algorithms and knowledge-aided STAP algorithms.Otherwise,it can better protect target signal and improve clutter suppression performance of knowledge-aided STAP algorithms.With using the second method,the covariance matrix reconstruction method based on temporal spectral estimation is improved,which reduces the error of covariance matrix reconstruction.Aiming at the off-grid problem of SR-STAP algorithm and the problem of heterogeneous clutter,an improved off-grid STAP method based on local search process is proposed,which significantly improves clutter suppression performance of short-range heterogeneous clutter.The main work of this thesis are summarized as follows:1.Amplitude and phase error estimation method is studied.Aiming at falling target gain and clutter model mismatch caused by amplitude and phase error of the array,two error estimation methods are studied.The first one makes the average output power of knowledge-aided clutter suppression as the cost function,which can accurately estimate error with iterative mode and multiple snapshots.We can jointly obtain the filtered output with using the way of integrating error estimation into clutter suppression processing.The second one is an iterative error estimation method with temporal reduced-dimension processing and multiple snapshots.It only uses the data of main-lobe clutter to estimate error,which can effectively reduce the calculation load.2.Covariance matrix reconstruction method based on temporal spectral estimation is studied.In heterogeneous clutter environment,it's hard to get enough i.i.d.samples to estimate covariance matrix.The covariance matrix reconstruction method based on temporal spectral estimation can effectively alleviate this problem and get good performance in the small sample case,but the model mismatch caused by error will result in performance degradation.With using the second error estimation method,an improved covariance matrix reconstruction method based on temporal spectral estimation is proposed,which can overcome the influence of the error and improve clutter suppression performance.3.Off-grid SR-STAP method is studied.The SR-STAP method usually uniformly divids the grid points in the angle-Doppler plane to build clutter basis,but the real clutter ridge may not just fall on these points,which is called off-grid problem.Considering the heterogeneous clutter caused by non-sidelooking airbrone radar and complex ground scattering characteristics,the low-complexity off-grid STAP algorithm based on local search clutter subspace estimation is improved.The improved method obtains sample data by using sub-aperture smoothing operation,and builds new cost function to select vectors.This improved method can achieve better clutter suppression performance with small sample or even single sample.