| As a basic problem of array signal processing,DOA(Direction of Arrival)estimation has a very wide range of areas in many fields,such as radar detection,underwater target location,etc.With the continuous development of marine resources,underwater DOA estimation plays an increasingly important role.Sound signal is commonly used in underwater exploration,and the traditional underwater DOA estimation algorithm regards acoustic velocity as a constant calculation.However,in the actual underwater acoustic environment,the acoustic velocity is constantly changing due to temperature,salinity and other factors.Considering the acoustic velocity as a constant will lead to larger estimation errors,so it is an urgent problem to overcome the influence of acoustic velocity on underwater DOA estimation.In the research of underwater DOA to overcome the influence of acoustic velocity,existing research obtains DOA estimation methods independent of acoustic velocity by adding arrays and utilizing geometric relations between arrays.But these methods require highly complex pairing algorithms.When the number of target signals increases,the estimation speed increases in a factorial order.In order to solve the problem of high complexity of existing methods,this paper studies the fast acoustic velocity-independent estimation algorithm based on arbitrary crossed linear array for one-dimensional and two-dimensional DOA estimation respectively.The algorithm proposed in this paper not only eliminates the influence of acoustic velocity in the estimation of the algorithm,but also is faster and more efficient than the existing acoustic veloc.The specific contents are summarized as follows:1.The theoretical basis of DOA estimation is introduced,including spatial spectrum estimation,array structure and mathematical model commonly used in DOA estimation,matrix algebra theory and so on.In addition,the performance indicators related to DOA estimation are also introduced.Simulation of one-dimensional and two-dimensional DOA estimation based on conventional ESPRIT algorithm is carried out with unknown acoustic velocity,and the influence of unknown acoustic velocity on the performance of conventional estimation is analyzed by simulation experiments.2.The existing DOA estimation algorithm which regards acoustic velocity as a constant is simulated,and the influence of acoustic velocity error on DOA estimation accuracy is analyzed.At the same time,the performance of the existing DOA estimation algorithm is simulated and the computational complexity of the algorithm is analyzed.According to the influence of unknown acoustic velocity on the performance of conventional two-dimensional DOA estimation,a two-dimensional DOA estimation algorithm based on adjustable angleuniform linear array is proposed,this algorithm can make full use of the stereoscopic geometry relationship between uniform linear array and eliminate the influence of acoustic velocity factor on the estimation performance of pitch angle.The simulation experiment results the proposed algorithm is superior to the conventional two-dimensional DOA estimation algorithm and the existing related two-dimensional algorithm in estimation performance of pitch angle with unknown acoustic velocity.3.For the case of one-dimensional DOA estimation,a fast estimation algorithm based on arbitrary crossed linear array is proposed.The algorithm uses the geometric relationship between two crossed linear arrays to derive one-dimensional DOA estimation expression which is independent of acoustic velocity and does not need parameter pairing.At the same time,the computational complexity of the algorithm is analyzed and its performance is simulated.4.The fast acoustic velocity-independent algorithm for one-dimensional DOA estimation is extended to two-dimensional DOA estimation.An array is added on the basis of the original array structure.According to the geometric relationship between arrays,the expression of two-dimensional DOA estimation with low pairing complexity independent of acoustic velocity is derived.Then the complexity of the two-dimensional DOA estimation is analyzed and its performance is analyzed by simulation. |
