基于共轭梯度的宽带相关处理快速算法  

Fast algorithm of wideband correlative processing based on conjugate gradient

作　　者：明星[1] 苑秉成[1] 刘建国[1] 

机构地区：[1]海军工程大学兵器工程系,湖北武汉430033

出　　处：《系统工程与电子技术》2010年第12期2533-2536,共4页Systems Engineering and Electronics

基　　金：海军"十一五"预研基金(4010703010102)资助课题

摘　　要：宽带主动信号处理是以宽带相关计算为基础的,通常采用目标尺度-时延二维搜索的计算结构。要达到较高的参数估计精度,处理结构就必须采用高密度搜索,但这同时会带来系统难以承受的大运算量。提出了基于共轭梯度的宽带相关处理快速算法,通过以较大的时延-尺度扫描间隔对目标参数进行粗略估计,得到参数估计的初值,并采用共轭梯度法构造一组共轭方向,由初值出发作迭代运算实现对峰值的搜索。仿真计算表明:基于共轭梯度的宽带相关处理快速算法能以较小的运算量得到对参数的精确估计,并具有较高的参数估计精度。参数估计误差的方差接近宽带相关处理的克拉美-罗下限(Crame-Rao low bound,CRLB)。Wideband active signal processing is realized by wideband correlative processing which has a time scale and time delay 2-D peak value search calculation structure.Aiming at the relatively high estimative precision,the structure must have a high search density,but it would cause a heavy operation capacity which system can not bear.Then the fast algorithm of wideband correlative processing based on conjugate gradient is advanced.The initial value is calculated by time delay and time-scale scanning process with wide scanning intervals.By the conjugate direction vector with conjugate gradient algorithm,the peak value is gained by iterative calculation which starts from the initial value.The result of simulation shows that the fast algorithm could get the peak value by low operation capacity and relatively high parameters estimative precision by iterative calculation.The variance of parameters estimation is close to the Crame-Rao low bound（CRLB） inferior limit of wideband correlative processing.

关 键 词：宽带信号处理 快速算法 共轭梯度法 参数估计 

分 类 号：TN911.7[电子电信—通信与信息系统]