Research On Robust Beamforming And Pattern Synthesis

Adptive beamforming and pattern synthesis play important roles in array signal processing. They are both extensively applied in radar, communication and acoustic diagnostics, geology and many other fields. Adaptive beamforming is designed to adaptively adjust weighted vector of the array according to the environment in real time so that the beamformer could get optimal performance under a certain optimality criterion. Pattern synthesis is applicated to make the actual pattern as much as possible to meet the desired pattern at the mainlobe shape and sidelobe height by the design of array parameters.Traditional adaptive beamforming algorithms have no robustness when there is a mismatch in the array parameter and environmental models. And this kind of mismatch will cause a serious deterioration in the performance of adaptive beamforming. To pattern synthesis, the problem is that some classic pattern synthesis algorithms cannot work for arrays with arbitrary element positions. And iterative algorithms based on adaptive theory usually converge slowly.Faced with the above problems, a mathematical model of array signal as well as its environment is established at first in this thesis. Then several adaptive beamforming algorithms are analyzed and compared under MSINR (Maximum Signal to Interference and Noise Ratio) criterion. Especially, the robustness of these algorithms are discussed. On the basis, three kinds of algorithms are studied, namely diagonal loading beamforming, eigen-subspace beamforming, and LCMV (Linear Constraint Minimum Variance) beamforming. In view of the drawbacks of traditional diagonal loading beamforming, we make further study of weighted vector norm restriction algorithms and steering vector uncertain set constraint algorithms. These algorithms bring commendable improvement of the beamformer output performance.The classic Chebyshev method is simple and efficient, but it cannot work for arrays with arbitrary element positions. To solve this problem, adaptive array theory is applied. The artificial interferers of various power levels are assigned in sidelobe regions to control sidelobe levels of the synthesized pattern. Then combining with the mainlobe shaping methods, a synthesized pattern after iteration can be achieved which is quite similar to the desired pattern. An improved method for conformal array pattern synthesis is proposed in this thesis. It’s based on LCMV criterion and is applicable to any array placement. Moreover, it can accelerate the convergence rate and reduce the dependence on the iteration coefficient for adaptive beamformers.