Adaptive 2-D DOA Estimation For Aribitrary Array Structures

Direction-of-arrival(DO A) estimation is a ubiquitous task in array processing. The conventional Multiple Signal C lassification(MUSIC) method is search-based and often computationally expensive, particularly in the application of joint azimuth and elevation estimation. In the thesis, we propose an adaptive 2-dimensional direction finding framework to track multiple moving targets for a rbitrary array structures by using the manifold separation technique(MST). First, we employ the subspace tracking technique to update the eigenbasis recursively on the arrival of a new data snapshot. In addition, a fast one-step operation is introduced to update the coefficient matrix in parallel by using the shift matrix. Finally, 2-D fast Fourier transform(FFT) algorithm is performed to compute the 2-D spatial spectrum at once. In comparison with the traditional MUSIC or root-MUSIC methods, the proposed method can reduce the computational complexity considerably, and be easy for hardware implementation.The thesis is summarized as follows:First of all, we briefly summarize the research progress of adaptive twodimensional DOA estimation, and analyze the structure model of antenna array. Then we study matrix models which the ideal and error conditions, coherent source and distributed source.Secondly, we also introduce the traditional DOA estimation algorithm, including the implementation steps of the classical MUSIC and root-MUSIC algorithm, and the comparison of advantages and disadvantages. And we study the improved algorithm of MST-based root-MUSIC in the one-dimensional situation: the fast root-MUSIC and IDFT-based root-MUSIC. Then there is also a comparison and analysis for two kinds of two-dimensional DOA algorithms.Finally, the core algorithm of this thesis is given: the adaptive two-dimensional DOA estimation algorithm based on FFT. The method makes the coefficient matrix divide into two parts. F irstly, the time- independent part only executes one time and can be obtained by off- line. Secondly, the time-related part uses the fast approximate iterative power(FAPI) method of the subspace tracking technology to obtain the weight matrix, and the shift matrix is introduced to make a two-step method become a single step parallel method for updating the coefficient matrix. Then the two-dimensional spatial spectrum is converted to the two-dimensional FFT operation to search the corresponding DOA estimated values quickly. Finally, the simulation analysis and comparison are also given to prove that the adaptive algorithm can not only reduce the computational complexity, but also can easily be used for hardware implementation.