Efficient reiterative censoring of robust STAP using the FRACTA algorithm

This paper presents further developments of the FRACTA algorithm (S.D. Blunt et al, NRL Tech. Rep. 10.059), which has been shown to be robust to nonhomogeneous environments containing outliers. The focus here is on the efficient implementation of the FRACTA algorithm. The key development is a censoring stopping mechanism whereby the number of reiterative steps can be minimized and computation is reduced. We introduce a data-dependent stopping rule that demonstrates excellent results as evidenced by the detection of targets in the KASSPER challenge data cube. We also present some other enhancements to the FRACTA algorithm that further improve both efficiency and performance.