The LFM Signals Detection And Parameter Estimation Base On FRET

Linear frequency modulation (LFM) signals are widely used in many informationsystems and fractional Fourier transform (FRFT) can detect LFM signals effectively. Forthe purpose of computational complexity reduction and multicomponent LFM detection,this thesis proposes several LFM signals detection algorithms based on FRFT. The maincontents are as follows.1. The basic definition, the attributes and the digital algorithms of FRFT areintroduced at first. Then, the theory of LFM signal detection based on FRFT is analyzedand the two-dimension peak search algorithm is given.2. In order to decrease the computational data volume, the relationship between theoptimal rotation angle of FRFT of the undersampled signal and that of the original signal isstudied. It is proved that the chirp-rate of noiseless LFM signal can be estimated correctlyeven though the signal is undersampled. By deducing and analyzing the signal to noiseratio (SNR) in the optimal fractional Fourier domain, the impact of sub-sampling rate ondetection of LFM signal is discussed and the rule of undersampled coefficient selection isproposed. A novel method is proposed to realize fast detection of LFM signal and theparameter estimation of LFM signal can be realized rapidly by applying FRFT to theundersampled signal.3. For the sake of decreasing operation time, discrete polynomial-phase transform(DPT) algorithm and BFGS (Broyden-Fletcher-Goldfarb-Shanno) algorithm are introducedand fast LFM signal detection algorithm based on DPT-BFGS are proposed. It candramatically reduce the operation time. Furthermore, the limitation of this algorithm isdiscussed.4. This thesis makes a further research on multicomponent LFM signals detectionbased on fractional autocorrelation. On the basis of a fractional autocorrelation definition, the relationship between multi-LFM signals shading coefficient and the LFM signalsparameters is deduced, and then a novel fast detection algorithm which has less datavolume is proposed. In addition, pretreatments which can restrain envelope detector’sglitches caused by noise are introduced to determine the threshold and enhance thedetection performance under high noise condition. Simulation results show that theproposed algorithm has sound detection accuracy under the low SNR condition and thedata volume is considerably small.