Modern Spectral Estimation Technology Research And Application Of FPGA Implementation

In the field of engineering application, analysis of the signal power spectrum can get signals on different frequency energy distribution, convenient for subsequent processing of signals. Spectrum estimation research all parameters of the signal in the frequency domain, the purpose is to extract useful signals submerged by noise, for power spectrum estimation method can be divided into classical spectrum estimation methods and the modern spectrum estimation method. Modern spectral estimation method has the characteristics of high resolution, and especially suitable for the short sequence data of spectrum estimation, so in communication, aerospace, radar, voice and image processing, and other fields have a large number of applications.This paper studied the general situation of the development and spectral estimation algorithm, spectrum estimation research significance were analyzed. Secondly on the basis of classical spectral estimation algorithm, and then some modern spectral estimation algorithm is studied, including the parameters of AR model spectrum estimation algorithm, MUSIC spectrum estimation algorithm, spectrum estimation algorithm and MVDR Burg spectrum estimation algorithm. The performance of each algorithm was introduced in detail as well as the advantages and disadvantages, and a software simulation is performed for each algorithm.Second, the detailed research on the basis of the existing algorithm, high computational complexity for MVDR algorithm and Burg algorithm in the estimation precision is poor, the characteristics of the improved method is proposed and is verified. Then, modern spectral estimation method in two-dimensional microwave imaging and noise source identification is verified in the scene. In microwave imaging scenario, the algorithm is extended to two-dimensional radar data, and the image data before using the improved algorithm and improved algorithm are analyzed and compared, in order to verify the validity of the improved algorithm; In noise source identification of a scene, and through the multiple source signals using the modern spectrum estimation method, estimate the noise power at each frequency point, analyze the main noise source and its energy distribution and frequency characteristics, which provide the basis for control of noise.Finally, the paper gives the improved algorithm on FPGA development board Virtex-5 implementation scheme design, to reduce resource consumption in the structure design and improve the processing speed is optimized. Through the implementation of hardware and the software simulation results contrast, verifies the correctness of the function of each module, and illustrates the algorithm’s feasibility in hardware implementation.