面向RCS序列周期估计的高斯过程优化Lomb-Scargle算法OA

Aiming at the problem of non-uniform sampling time of radar cross section(RCS)sequence caused by the limitation of tracking resources in phased array radar,this paper proposes a period estimation meth-od using a Lomb-Scargle algorithm optimized by Gaussian process.In this method,the Lomb-Scargle algorithm is first adopted to process the non-uniform sampling sequence and obtain robust periodic initial value;then,a Gauss-ian process model including the periodic kernel is constructed,and finally,this initial value is taken as the starting point to iteratively optimize the period and other hyperparameters through the log-likelihood function.Simulation results demonstrate that,under conditions of a low proportion of observation duration and a low signal-to-noise ra-tio,the proposed method achieves fewer root mean square error than traditional trigonometric function fitting,sig-nificantly improving the period estimation accuracy of non-uniformly sampled RCS sequences.