Matrix Decomposition Methods for Efficient Hardware Implementation of DOA Estimation Algorithms: A Performance Comparison

Matrix operations form the core of array signal processing algorithms such as those required for direction of arrival (DOA) angle estimation of radio frequency signals incident on an antenna array. In this paper, we present a performance comparison of matrix decomposition methods for efficient FPGA hardware implementation of DOA estimation algorithms. These methods are very important in subspace-based DOA estimation algorithms as they are used for signal space extraction. DOA estimation algorithms employing LU, LDL, Cholesky, and QR decomposition methods are implemented on a Xilinx Virtex-5 FPGA. These DOA estimation algorithms are simulated in LabVIEW as well as experimentally validated in real-time on a prototype testbed constructed using Universal Software Radio Peripheral (USRP) Software Defined Radio (SDR) platform from National Instruments. Performance comparison of these algorithms is made in terms of resources consumption, computation speed, and estimation accuracy.