Search

An FPGA implementation and real-time experimental verification of proposed direction of arrival (DOA) estimation algorithm employing LDL factorization are\ud presented in this paper. The proposed algorithm is implemented on a Xilinx FPGA using LabVIEW software and its real-time experimental verification is performed using National Instruments (NI) PXI platform. The proposed method has several advantages over well-known methods which are based on either eigenvalue decomposition (EVD) or singular value decomposition (SVD). It provides faster\ud execution since LDL factorization requires O(n3/6) number of operations whereas EVD requires O(n3). Results from Matlab simulations and real-time experiments demonstrate the effectiveness of the proposed method. Successful FPGA\ud compilation reports show low resource usage and faster computation time for LDL-based method compared with QRbased implementations. Performance comparison is done in terms of estimation accuracy, FPGA processing time and\ud resource utilization.

Estimating the channel state information (CSI) of several transmitters that use orthogonal space-time block codes (OSTBC) to communicate with a single receiver is considered. Based on Rank Revealing QR (RRQR) factorization and Propagator method (PM), two new algorithms to estimate multiuser MIMO channels are proposed. The algorithm estimates the subspace spanned by the user channels use only a few training blocks to extract the users CSI from this subspace. Our both algorithms achieve better performance in comparison with reference LS based approach. Computer simulations are included to validate proposed methods.

In this paper, we proposed a closed form solution for blind Orthogonal Frequency Division Multiplexing (OFDM) Carrier Frequency Offset (CFO) estimation employing the Rank- Revealing QR triangular factorization Method (RRQR). The advantage of using the RRQR it gives precious information about numerical rank and efficiently separates the signal space from the noise space. Furthermore, the RRQR does not involve the eigenvalue decomposition (EVD) or singular value decomposition (SVD) of co-variance matrix of the received signals and is a valuable tool in linear algebra to meet compatibility requirement of real time environment. Computer simulations shows the superior performance and much less processing time of RRQR compared with the method employing ESPRIT Algorithm Index-Terms: OFDM, Carrier Frequency Offset, DFT, Estimation, RRQR Method.

Joint Time Delay and Frequency Estimation (JTDFE) problem of complex sinusoidal signals received at two separated sensors is an attractive problem that has been studied for many engineering applications. In this paper, the Rank-Revealing QR factorization is applied to the real data matrix obtained via the unitary transformation of the square Toeplitz complex data matrix. Then the MUSIC spectrum estimation function is used to estimate the frequencies. The time delay is estimated by applying RRQR to the complex data matrix. The unitary transformation from complex to real would reduce the processing time of frequency estimation by almost a factor of four, since the cost of complex manipulations is four times the real manipulations. Also RRQR is an important tool in numerical linear algebra because it provides accurate information about rank and numerical null space. The simulation results validate the performance of the proposed method.