RCS Series Generation of Targets with Uncertainty Using Gaussian Process

Radar cross section (RCS) is one of a target's radar signatures, and its series is closely related to the generation of radar signal. Nowadays researchers usually set RCS as deterministic values or a random variable in radar signal simulation. However, the precise RCS value of many targets cannot be acquired due to the uncertainty in modeling the RCS of targets. Furthermore, the RCS values of different angles may not be independent, so the correlation between them should be considered. In this work, we propose an RCS series generation method for targets with uncertainty using correlated Gaussian process. The first step is to compute the covariance matrix of RCS data in different view angles. Then is to do the eigen decompostion of the covariance matrix, and get the multivariate Gaussian random variables as the generated RCS series. Numerical examples have validated of such method.