The Super-Resolution DOA Estimating Algorithm Based On Arrays Of Small And Medium Aperture

As the techniques employed in mobile communication, remote control and measure-ment, and aerospace engineering are developing rapidly, the methods to passively detectthe direction and other parameters of the targets by the their radiated magnetic signalsreceive more and more attention. Arrays of small and medium aperture which can beloaded on moving facility, suchas ships, vehicles, aeroplanes and satellite, become themain platform of the passive detection technique application. In recent years, the devel-opment in electronics has made it possible to integrate the whole system of antenna array,microwave equipment and back end processing device in to a small space, with smallvolume and little weight. Mismatching with the powerful property of the hardware, thepassive detection algorithms practically used are of limited accuracy and unable to resolvemultiple signals. The subspace DOA estimating algorithms suchas MUSIC and ESPRIT,possessing excellent accuracy performance approaching to Cramer-Rao bound, and thecapacity to resolve close signals, are proposed in last decades. These algorithms are suit-able to be applied in arrays of small and medium aperture. However, because of someproblems encountered in practical application, they are not widely used in actual systems.The thesis did some research to some critical problems in actual systems, provided solu-tions to them, and designed an experimental system running MUSIC-like algorithms on asmall size array.Many narrow-band signals of complete polarization exist in practical environment.Using the polarization message of these signals to enhance the accuracy and resolutionperformance of algorithms will be very helpful to passive DOA estimation system. Thethesis extended MUSIC algorithm to polarization sensitive array, and analyze propertiesof MUSIC’s space spectrum under such condition. Based on the properties, we proposed atechnique to alleviate the computation complexity invoked in the algorithm. The enhancedperformance in accuracy and resolution are analyzed. Monte-Carlo simulation resultsare provided to illustrate the algorithm’s advantage in performance, and the complexityin computation are given. Both theoretical analysis and simulation results support thealgorithm’s validity.Although subspace algorithms have presented excellent performances in DOA esti-mation, the factors of array must be considered to improve the design of the array, in orderto make the algorithm to function efficiently. The factors of array mainly include the ar-rangement of the array and the pattern characteristics of the individual element. Similarto traditional DBF algorithm, MUSIC and other subspace algorithms face the problemof ambiguity phenomenon. To avoid the ambiguity by increasing the density of arrayelements will boost the system complexity. Arrays of large aperture may make peaks inMUSIC space spectrum too sharp, and require intensive sampling of the spectrum to lo-cate the peaks’ positions, which renders a disadvantage in practical application. An arraycombined of subarrays with various apertures is designed. The inner subarray of smallaperture can help to elude the array ambiguity, and the outer subarray of large apertureenhances the performance of algorithm. Some adjustments are made for ordinary MU-SIC to reduce its complexity. Concerning the pattern of the array element, we suggestthat the characteristics of element’s pattern can be explored to improve the performanceof MUSIC-like algorithms. Base on the amended MUSIC algorithm which adapts to thearray element with different patterns, the validity of this idea is verified, and its feasibilityis further confirmed by Monte-Carlo simulation.One weakness of subspace algorithms lies in the difficulty of estimating coherentsignals’ bearings. In real practices, multipath interferences, active radar decoy, and etc.,may generate multiple coherent signals. Estimation of these signals’ directions become anurgent requirement of practical application of subspace algorithms. As arrays of mediumor small aperture are often loaded on moving platforms, enlightened by the spatial s-moothing algorithms, we presented several methods to detect coherent signals, accordingto platforms of different velocities. Detailed analysis are proposed in cases of high-speedand curvingly moving platforms, to make sure whether the method can detect coherentsignals, depending on the combination of velocity, aperture, SNR and snapshot number.Simulation results are provided to illustrate the validity of the methods.To verify the feasibility of MUSIC and MUSIC-like algorithms on arrays of smalland medium apertures, we have designed and realized a experimental system running fastpolarization MUSIC algorithm. The system adopts a Uniform Circular Array of smallsize, to sample and estimate the direction of signals within the 1-2GHz frequency range.This experiment achieved favorable results in signals’ directions estimation.