Research Of DOA Estimation Subspace Algorithm Of Wideband Signal

Signal DOA (Direction of Arrival) estimate algorithm is in the array signal processing domain a hot spot's research problem, mainly solves in the space many signal precise direction finding questions. This article unifies the **** research, mainly carries on the wideband signal DOA estimate the subspace algorithm research.First, analyzed through the simulation experiment affected the DOA estimate result in the practical application the important attribute, and to classics MUSIC algorithm (Multiple Signal Classification) and ESPRIT algorithm (Estimation of Signal Parameters via Rotational Invariance Techniques) the theoretical performance and good and bad points carried on the contrast, the MUSIC algorithm belonged to the noise subspace algorithm, obtained the signal incidence direction using the guidance vector and noise subspace's orthogonality, the algorithm performance is slightly good in the ESPRIT algorithm; The ESPRIT algorithm belongs to the signal subspace algorithm, uses among the submatrices the signal subspace revolving invariable characteristic to solve, but the ESPRIT algorithm does not need to carry on the spectrum peak search, timeliness is good.Next, the precision which exists in view of the ISM algorithm is low and computation load big these two spots proposed two improvement algorithms insufficient:First, from enhancing accurate angle makes weighted processing to the signal energy; Second, carries on the improvement from reduced computation load's angle to the signal structure. Through the simulation experiment, has confirmed two improvement algorithm superiority from root-mean-square error and resolution probability.Finally, in the system profoundly has studied in the classics CSM kind of algorithm foundation, discussed the focusing matrix's structure method and focusing frequency's selection criterion, and to not needing the angle estimated the wideband signal DOA estimate method to carry on the simulation analysis and improvement research. The result indicated that after improvement algorithm already does not need the angle to estimate, and around focuses does not have the signal-to-noise ratio to lose, has a better performance.