Research on High-Efficient Spatial Spectrum Estimation Algorithm
|School||Harbin Institute of Technology|
|Course||Information and Communication Engineering|
|Keywords||Spatial spectrum estimation MUSIC algorithm ESPRIT algorithm Single Snapshot Parallel processing|
Spatial spectrum estimation technique used to estimate the direction of arrival of the signal processing bandwidth, research to improve the angular resolution and the estimated accuracy of the algorithm, as well as to improve the processing speed of the algorithm. The traditional algorithm constrained by the Rayleigh limit, lower angular resolution, then there superresolution algorithms breakthrough this limitation, and has the characteristics of high-resolution and high estimation accuracy. MUSIC and ESPRIT algorithm is the most typical super-resolution algorithm, these two algorithms proposed greatly promoted the development of the spatial spectrum estimation algorithm. However, these two algorithms exist in the practical application of computational complexity is not suitable for real-time processing, the estimated accuracy is not high, and under the conditions of a single snapshot, efficient spatial spectrum estimation algorithm to solve these problems, came into being. Efficient space spectrum estimation algorithm to reduce the amount of calculation and improve computing speed and improve the algorithm estimation accuracy aspects expand. First, this paper introduces the basic principles of the MUSIC algorithm, the computational complexity for the algorithm is not suitable for real-time processing, a parallel processing program. Op conversion pretreatment by the real value to the real number field, and then use the Householder transform the original covariance matrix into tridiagonal matrix and its QR decomposition, and finally on the various stages of a multi-processor parallel processing. Then, the parallel algorithm extended to the ESPRIT algorithm. In this paper, Lanczos transform the non-symmetric covariance matrix into a tridiagonal matrix, and then use the QR algorithm with origin displacement characteristic decomposition constructed covariance matrix eigenvalue decomposition are suitable for parallel processing. Finally, for the conditions of small number of snapshots, the estimated accuracy of the MUSIC algorithm to reduce the problem, this paper studies a Toeplitz matrix based the dimensionality reduction MUSIC algorithm and its improvements. Receive data to the signal subspace projection preprocessing to obtain new data dimensionality reduction method to estimate the covariance matrix, re-use basic MUSIC algorithm for DOA estimation. Dimensionality reduction algorithm, however, will make the antenna degrees of freedom to reduce and the a Toeplitz matrix dimensionality reduction MUSIC algorithm, through the Toeplitz features to construct the covariance matrix, not loss of antenna degrees of freedom is proposed to solve this problem. Simulation results show that the MUSIC and ESPRIT algorithm parallelization little effect on the performance of the algorithm, which greatly reduces the computational complexity of the algorithm to improve the processing speed. Single snapshot conditions proposed Toeplitz matrix based MUSIC algorithm can improve the estimation accuracy of the algorithm performance, especially in the low SNR good performance, and can effectively estimate coherent source.