Research On Array Pattern Synthesis And Adaptive Beamforming Technologies

Array pattern synthesis and adaptive beamforming have found a wide range ofapplications in radar, sonar, wireless communication and astronomy, et.al. By thepattern synthesis or adaptive beamforming techniques, the systems in these applicationscan transmit/receive the signal to/from a specified direction or a range of directions, andtransmit/receive as little signal as possible to/from other directions, or completelysuppress the signal to/from some directions. However, in some kind of arrays or in somespecially situations, the exsiting methods will suffer from some kind of problems. Thisdissertation summarizes the applications of some existing pattern synthesis methods andproposes some modified methods, and then focuses on the pattern synthesis ofconformal arrays and large arrays. New pattern synthesis methods for conformal arraysand large arrays are proposed to address the shortcomings of the conventional methods.To improve the robustness of the adaptive beamforming against model errors, severalnew robust beamformer are presented. The main contributions are illustrated as follows:Firstly, the dissertation summarizes the applications of the convex optimization onpattern synthesis of pencil beam or shaped beam patterns, and a shaped beam patternsynthesis method with nonuniform arrays is proposed and is applied to array failurecorrection. A pattern synthesis method based on convex optimization is proposed tosynthesize the sum and difference patterns. This method can optimize the elementpositions and excitations for sum and difference patterns simultaneously. Therefore, thesystem complexity can be reduced. The dissertation introduced the particle swarmoptimization method, and proposed a new fuzzy discrete particle swarm optimizationmethod. What’s more, the matrix pencil method based pattern synthesis method isintroduced and extended to2D arrays.Secondly, the pattern synthesis of conformal arrays is considered. The first topic ofthis part is to optimize the array excitations of a given array to obtain a desired pattern. Iproposed a novel active element pattern decomposition based method of conformalarray synthesis. This method requires less pattern samples, less memory, lesscomputational cost and can steer the mainlobe or nulls to precise directions. Then, anoptimization method on conformal array element positions is proposed. In this method,the pattern of an element located at an arbitrary position on the platform is obtained by interpolating the element patterns at several sampled locations, and the mutual couplingbetween elements is considered by robust pattern synthesis method. The full wavesimulation results validate these two proposed methods.Thirdly, this dissertation deals with the fast pattern synthesis method for largearrays. Inspired by the nonuniform fast Fourier transform, I proposed a novel fast pencilbeam pattern synthesis method for large unequally spaced antenna arrays. This methodis based on fast Fourier transform, so larger arrays can be handled and can obtain betterresults when compared with other methods. The dissertation also considered the flat-topfootprint pattern synthesis in this part. The dissertation proposed a synthesis method thatcan obtain arbitrary footprint pattern which is important in satellite applications. Theproposed method chooses the antennas in the given uniform array and the dynamicrange ratio can be controlled.Finally, the dissertation works on the robust adaptive beamforming methods. Thedisseratation proposed an adaptive beamforming method that is robust against largearray calibration errors. This method reconstructs the interference covariance matrix andestimates the steering vector associated with the desired signal. Compared withstate-of-the-art methods, the proposed method can obtain better results even when largecalibration error exists. What’s more, the dissertation transform the adaptivebeamforming problem into a weighted array pattern synthesis problem. The robustnessof the adaptive beamformer is improved by adding restriction on the mainlobe widthand sidelobe levels. The proposed method can suppress the interferences with highpriority, and it is robust against model errors.