Unrestricted.2016-11-142011-02-182016-11-141998-08http://hdl.handle.net/2346/10042The output of space-time adaptive processing (STAP) is a weighted sum of multiple radar returns, where the weights for each return in the sum are calculated adaptively and in real-time. The most computationally intensive portion of most STAP approaches is the calculation of the adaptive weight values. Calculation of the weights involves solving a set of linear equations based on an estimate of the covariance matrix associated with the radar retum data. The traditional approach for computing the adaptive weights is based on a direct method called QR-decomposition. This method has a fixed computational complexity, which depends on the size of the equation matrix and provides the exact solution. An alternative approach based on an iteractive method called Conjugate Gradient is proposed, which allows for trading off accuracy for reduced computational complexity. The two approaches are analyzed and compared. Existing computational strategies for STAP typically rely on the use of multiple digital signal processors (DSPs) or general-purpose processors (GPPs). An alternative strategy is proposed, which makes use of Field Programmable Gate Arrays (FPGAs) as vector co-processors that perform inner product calculations. Two different "innerproduct co-processor" designs are introduced for use with a host DSP or GPP. The first has a multiply-and accumulate structure and the second uses a reduction-style tree structure having two multipliers and an adder. The proposed strategies are implemented and compared to the traditional method.application/pdfengAlgorithmsAdaptive filtersSignal processingRadarAdaptive signal processingThesis