Implementation of QRD_MGS algorithm for space-time adaptive processing based on FPGAs

Space Time Adaptive Processing (STAP) can curbthe effects of interference significantly. However, the calculationof STAP weights, including QR decomposition (QRD) andsolving linear equations needs drastic computation. This paper,focuses on improving the QRD algorithm by modified gramschmidt(QRD_MGS) method and presents an efficient FPGAdesign by adding a parameter called vector size that, enables theimplementation of multi-vector. In this work, the structure ofQRD_MGS algorithm is adjusted for effective floating pointimplementation, and the design is described in detail. Theresource utilization, operating frequency, power consumption,delay and throughput measured and reported for several matrixsizes and vector sizes. The results show that the proposed methodcan satisfy the requirement of real-time computation of adaptiveweights and to be used for implement the full system of STAP.

کلیدواژه ها:

FPGA ، QRD_MGS ، Implementation ، STAP ، ULA