DOA Estimation Of LFM Sources Based On Fractional Fourier Transform

The array direction finding technology plays a very important role in many aspects ofdaily life and works, from private mobile communications to national defense. Direction ofArrival(DOA) estimation is widely used as one of important methods of array directionfinding technology. The technology plays a pivotal role in the development of nationaldefense and security and economic development, such as the anti-interference ability of theradar scanning in the military field, the missile positioning accuracy, the reliability of medicalimaging in the economic field, the accuracy of seismic exploration, etc. In recent years, theLinear Frequency Modulation (LFM) signal as a typical non-stationary signal is paid moreattentions. The Fractional Fourier transform (FRFT) is a newly developed time-frequencyanalysis tool, it has more properties which are not be possessed by any other time-frequencyanalysis in the processing and analyzing of LFM signal. Therefore, the DOA estimation ofLFM signal based on FRFT is analyzed in this thesis.Firstly, introducing the traditional DOA estimation algorithm such as MUSIC algorithmand spatial smoothing algorithm, and applying them to the Non-uniform Linear Array(NULA), and then comparing the level of resolution of the two different array models, andanalyzing the influence on the resolution caused by the changes of different parameters.Secondly, introducing briefly the basic definition and the properties of FRFT, and thenresearching on the characteristics of the white Gaussian noise in the fractional domain and thequestion of dimensional normalization in the digital computation of the Fractional FourierTransform.Finally, combining FRFT with MUSIC to solve the DOA problems of the one-dimensional wideband LFM signals under the non-uniform linear array. Through thesimulation, the impacts on the resolution caused by the various factors are analyzedcomprehensively. It is proved that the resolution of NULA is improved and it is higher thanthat of ULA (Uniform Linear Array), at the same time it is closer to the actual engineeringapplications. Because the coherent signal source can not be handled efficiently by MUSIC inFRFT domain, spatial smoothing technology is adopted to solve the DOA estimation of thewideband LFM signals which have the same time-frequency features. Through the rationalarrangement of the NULA, the algorithm improves the resolution and practicability.