Research On Reduced-Dimension Space-Time Adaptive Processing Method And Clutter Pre-Filter Technique In Airborne Radar

The airborne early warning(AEW) radar that can move swiftly and scan broadly plays a very important role in the modern war. However, the airborne radar faces a more complicated clutter environment than the ground-based one due to its rapid movement. The space-time adaptive processing(STAP) method that is capable of suppressing the clutter in both the time domain and the spatial domain is an effective tool in airborne radar for suppressing the clutter and detecting targets. But some problems such as large computational-cost and high training sample demanding limit extensive applications of STAP and motivate the development of sub-optimal reduced-dimension and reduced-rank STAP algorithms. Up to now, there is no one reduced-dimension or reduced-rank STAP algorithm can be applied to the practical environment since training samples are intrinsically insufficient in real clutter environment. Therefore, the noval reduced-dimension STAP methods that further decrease the computational-cost and training sample requirement are studied in this dissertation.Pre-filtering the clutter before the adaptive processing will indeed diminish the clutter degrees of freedom(Do Fs) and enhance the performance of the following STAP methods. Therefore, according to the airborne clutter model, the non-adaptive clutter pre-filter that is low computational-cost and immune to the number of training samples is also studied in this dissertation. The main contributions of this thesis are summarized as follows: 1. The traditional post-Doppler adaptive beam-forming approaches such as factored approach(FA) and extended factored approach(EFA) can significantly reduce the computation-cost and training sample requirement in adaptive processing. However, their clutter suppression ability can be considerably degraded with the increasing number of antenna elements. Aiming at this problem, a two-stage reduced-dimension adaptive processing method based on the decomposition of spatial data is proposed. This method decomposes the spatial data after Doppler filtered into a Kronecker product of two short vectors. Then a bi-quadratic cost function is obtained. The circular iteration is applied to solve the desired weight. Experimental results show that the proposed method has the advantages of fast convergence and low computational-cost. It has greater clutter suppression ability especially in small training samples support compared to FA and EFA. 2. In recent years, multiple-input multiple-output(MIMO) radar has been receiving increasing attentions from researchers and engineers for its improvement in array aperture, angle resolution and Do Fs. However, the conventional STAP methods that applied in airborne MIMO radar will bear huge computational-cost and excessive training sample requirement for the increasing Do Fs especially in large-scale array antennas. Therefore, in order to making the STAP workable in airborne MIMO radar, a two-stage dimension-reduced adaptive processing method based on the spatial domain decomposition is proposed. Firstly, this method decomposes the weight vector in FA or EFA into a Kronecker product of two shorter weight vectors. Secondly, the original cost function of FA or EFA is turned into a bi-quadratic cost function. Thirdly, the cyclic minimizer is applied to find the desired weight vectors. Especially, since this method make the adaptive weight vector in FA or EFA become two considerably smaller weight vectors, it has much faster convergence rate and less computational complexity. 3. The two-dimensional pulse-to-pulse canceller(TDPC) of ground clutter can effectively pre-filter the clutter along the clutter trace. The moving target detecability of following space-time adaptive processing algorithms can also be enhanced after TDPC as the pre-filter. It can be applied in not only the side-looking airborne radar but also the non-sidelooking airborne radar. However, the inaccurate radar system parameters will degrade the TDPC’s clutter suppression ability. Aiming at this problem, a robust two-dimensional pulse-to-pulse canceller(RTDPC) of ground clutter is proposed. The errors of radar operating parameters and airborne velocity are taken as the priori knowledge and added into the design of TDPC filtering coefficient matrix. The simulated and MCARM real data are utilized to verify the clutter suppression performance of RTDPC with inaccurate airborne velocity and drift angle. The moving target detecability of following space-time adaptive processing algorithms are also enhanced. Our proposed RTDPC further extends the application range of the TDPC for the addition of parameters error in the clutter filter design. 4. Due to the complex geometry configuration of the receiving and transmitting airborne platform, the clutter of the bistatic airborne radar is strongly dependent on range. As a consequence, the performance of space-time adaptive processing(STAP) will be degraded for the great estimation error of the clutter covariance matrix. To solve this problem, a clutter pre-filtering method applied in the bistatic airborne radar that takes advantage of radar operating parameters, platform velocity and so on is proposed. The velocity error of the airborne platform is also considered. A majority of the clutter can be filtered so that the minority of the residual clutter can be completely suppressed by the well-developed STAP algorithm. The computer simulation results show that this method is effectively workable to several classical geometry configuration of bistatic airborne radar. The moving target detectability of the following STAP algorithm is also enhanced after this pre-filter. 5. The clutter covariance matrix is relatively low-rank compared with the degrees of freedom of the processor in space-time adaptive processing(STAP). Therefore we propose the clutter canceller based on the low rank of clutter subspace(LRCC) to suppress the ground clutter. This method constructs the original clutter by a relatively small quantity of linearly independent space-time steering vectors. Then clutter echoes received in adjacent pulses are cancelled to build our clutter canceller. The clutter canceller can be cascaded with the conventional spatial-temporal matching or STAP to enhance the following algorithms’ performance. Experiments results show the validity of this clutter canceller. Using our clutter canceller as the pre-filter, the following spatial-temporal matching or STAP will definitely outperform the original one.