Fpga, Signal Subspace

Array processing plays a very important role in many fields such as mobile comunication ,electronic reconnaissance, parameter estimation, signal discrimination and so on.Subspace-based method is one of the important methods of array processing and has producted plenty of high performance algorithms through years of development.The typical ones are MUSIC (Multiple Signal Classification), ESPRIT(Estimation of Signal Parameters via Rotational Invariance Technique) .Among all the algorithms based on subspace-based method, MUSIC proposed by R.O.Schmidt is high precise and high resolution. Nevertheless, it needs huge quantity of computation, which baffles its further application in realtime processing. Therefore, it is very necessary to study its high-speed implement methods.This dissertation studies the high-speed way to apply MUSIC algorithm in an four-sensor uniform linear array (ULA) system.(1) Analyzing the disadvantages of the present algorithm based on FPGA for eigen-decomposition, the dissertation proposes the MC-Jacobi algorithm which is better for eigen-decomposition when using FPGA. MC-Jacobi adds the parall of computation,so the time will be shorten by 2/3 comparing with the old algorithm.(2) Basing on MC-Jacobi algorithm,the dissertation designs EVD hardware diagram.The dissertation does function simulation and timing simulation after coding by VHDL which is one of description language for RTL. Making a model for real testing, the dissertation applys MC-Jacobi algorithm into FPGA on the demo board.(3) The dissertation designs the hardware diagrams of Corelation Matrix Module and Spectrum Peak Finding Module.Accroding to the hardware diagrams, the dissertation codes by VHDL and do function simulation and timing simulation.