将信号稀疏分解——正交匹配追踪(Orthogonal Matching Pursuit,OMP)引入到阵列信号处理领域,在OMP分解的基础上,提出了宽带Chirp信号多参数估计方法。首先根据宽带Chirp信号形式建立过完备原子库,对阵列接收信号在该过完备原子库上利用OMP做稀疏分解,从而由最佳匹配原子的参数获得信号的起始频率和调频斜率的估计,得到宽带Chirp信号形式。在此基础上,再根据阵列结构和已获得的宽带Chirp信号形式建立另一个原子库,通过计算阵列接收数据与原子库中原子之间的互相关矩阵的迹,搜索迹的最大峰值找出最匹配的原子,进而由最佳原子的参数获得信号的波达方向角度(Direction of Arrival,DOA)的估计。仿真实验证明了该算法对参数估计的有效性,并且表明与WVD(Wigner-Ville Distribution)方法相比,该方法能更有效地对信号的波达方向角度进行估计。
Sparse decomposition orthogonal matching pursuit (OMP) is applied to array signal processing field, and a new method for multi - parameter estimation of wideband Chirp signals based on OMP decomposition is presented. At first, according to wideband Chirp signal geometry, the over-complete dictionary is established and OMP decomposition is performed to the array receiving signal on the over-complete dictionary. According to the parameter of the best match atom, the estimation of starting frequency and modulating frequency of the signal is obtained, and then the wideband Chirp signal geometry is get. On this base, according to the array signal structure and the wideband Chirp signal geometry obtained, another over-complete dictionary is established. By calculating the trace of cross-correlation matrix between the array receiving data and the atom in the atom dictionary, and searching the maximum peak value, the best match atom is obtained. Then, according to the parameter of the best atom, the direction of arrival(DOA) estimation of the signal is realized. Simulation resuhs demonstrate that the algorithm is efficient to the parameter estimation and can estimate the signal's DOA more efficiently compared with Wigner-Ville distribution(WVD) algorithm.