Application Of Sparse Decomposition In Spatial Spectrum Estimation

Spatial spectrum estimation is an important area in array signal processing, which is widely used in radar, communication, sonar, earthquake, chronometer and other aspects of science and technology. The main purpose of spatial spectrum estimation is to estimate the parameters and source bearing of spatial signals using spatially displaced sensor arrays. It has become a hotspot in the field of array signal processing, and has obtained rapid development in both theoretical research and engineering practice, and wide development space in many military and civil domains.This thesis focuses on application of signal sparse decomposition in spatial spectrum estimation. As a new approach of signal decomposition and expression, signal sparse decomposition has been widely used in many fields of signal processing including signal compression and coding, signal identification, signal time-frequency distribution, etc. First of all, this thesis analyses the development and research status of spatial spectrum estimation, then introduces basic knowledge of spatial spectrum estimation and signal sparse decomposition. In the following chapters applications of signal sparse decomposition in spatial spectrum estimation are introduced respectively. In the third chapter, a measure of frequency estimation based on signal sparse decomposition is introduced. A frequency estimation algorithm based on MP decomposition is brought forward, and the performance of new algorithm well outperforms that of tradition DFT algorithm in low SNR situation. In addition, a method of signal time-frequency decomposition based on sparse decomposition with time-frequency atom is discussed and time-frequency analysis of non-stationary signal is obtained. Moreover, a new method of instantaneous frequency estimation is proposed. Chapter four presents application of signal sparse decomposition in DOA estimation. It starts with sparse decomposition of array signal, establishes proper over-completed atom dictionary, decomposes array signal adaptively on subspaces of each direction, and implements high-resolution estimation of DOA. A new algorithm of DOA estimation based on MP decomposition is proposed, and simulation proves the...