Research On Angle Estimation Techniques For MIMO Radar Based On The Compressive Sensing Theory

Multiple-input Multiple-output (MIMO) radar, as a type of new system radar, has higher resolution, lower intercept and capture probability and more freedom of degree, compared to the conventional phased-array radar. In the past few years, MIMO radar has drawn domestic and foreign scholars' considerable attention. Angle estimation is an important attribute of array signal processing, which cannot only achieve targets detection,but have a wide range of applications in radar, sonar, communication and so on. Continuous development of compressive sensing (CS) theory has been a heat research spot in the field of radar.Consequently, domestic and foreign scholars start to deal with problems in the traditional subspace-based methods by sparse signal representation (SSR) approach, looking forward to improve the angle estimation accuracy.From the perspective of reducing computational complexity and improving angle estimation performance, this paper is mainly focus on the research of angle estimation methods based on sparse representation scheme in monostatic MIMO radar. The main contents are listed as follows:Firstly, we consider both fundamental principles and framework of the CS theory, and analyze sparse signal in both spatio-temporal and transform domains, and sparse reconstruction conditions. Based on a scheme of sparse representation, the model of angle estimation-based compressive sensing is introduced. Then, two classes of subspace-based angle estimation methods, include spatial angle searching-based methods and rotation invariance subspace-based methods, are investigated. The spatial angle searching-based method includes reduced-dimensional Capon algorithm and the rotation invariance subspace-based methods include reduced-dimensional ESPRIT and conjugate ESPRIT algorithms. Lastly, the pros and cons of those traditional angle estimation algorithms are compared and summarized by simulation.Secondly, in order to increase the number of correct target resolution and enhance angle estimation accuracy, the characteristic of noncircular signal source is used. After analyzing the definition and mathematic model of noncircular signal, the model of angle estimation-based noncircular signal in monostatic MIMO is constructed. Then compressive sensing-based angle estimation for noncircular sources in MIMO radar is proposed. Because of considering the characteristic of noncircular structure, the virtual array aperture of radar is efficiently extended, so the proposed algorithm obtains better angle estimation performance.Finally, we consider of applying reweighted matrix and sparse representation theory into monostatic MIMO radar. The model of sparse signal representation in monostatic MIMO radar is reconstructed and then we proposed a reweighted sparse representation based angle estimation algorithm for monostatic MIMO radar. The proposed method can efficiently enhance the sparse solution. However, due to the specificity of reconstructing reweighted matrix, we cannot make some reduced-dimensional processing of received data in advance,which may result in higher computational complexity. In terms of this problem, on the other hand, a real-value reweighted sparse representation algorithm-based data reconstruction for angle estimation in monostatic MIMO radar is proposed. The signal model-based data reconstruction of monostatic MIMO radar is presented, then the central symmetry property of virtual array by unitary transformation is used to make a real-value operation for received data after data reconstruction. Therefore,a sparse representation scheme for angle estimation is proposed, which does better make the sparsest solution of l1 norm penalty approximate to the l0 norm penalty, as a result, efficiently improve the angle estimation performance.