Study On Array Calibration And Super Resolution On Angle Measurement For Hybrid Sky-surface Wave Radar

Due to its unique mode of transmission, high frequency hybrid sky-surfacewave radar can make an over-the-horizon detection and monitoring over theinterested sea area, it’s can provide an earlier vigilance time and manage air and seatraffic with high defense, political, military and economic value. HF hybridsky-surface wave radar’s transmitter and receiver placed in two different location s,so it can be classified as bistatic radar system. The receiver receives echoes silentlywith better invisibility and high viability. The radar system is relatively new withfew references. This thesis focuses on the HF hybrid sky-surface wave radar arraycalibration technique and super-resolution with single snapshot algorithms.This thesis’s research will focus on shored-based radar system. Firstly it willanalysis the HF hybrid sky-surface wave radar research status sketch it’s schematicdiagram, deduce the linear frequency-modulated continuous wave(LFCW) signalprocessing in HF radar. From this it got the range-doppler2-dimension processingprocedure. Added by the adaptive beamforming form the classicrange-doppler-azimuth processing procedure. Finally the thesis point out the singlesnapshot problem in the super-resolution algorithm followed-up. On the arraycalibration part of this thesis, first it describes the mathematical model of the idealarray signal and amplitude and phase error. Introduce the advantages anddisadvantages of active calibration and self-calibration, and principle of two classicself-calibration algorithm and FFT algorithm to be compared. It briefly analysis theself-calibration source can be used by bistatic radar, pointed out that the direct wavewith a high signal to noise ratio (SNR), strong orientation can be an idealcalibration source. After then utilize Monte Carlo simulation method, evaluate theperformance of Toeplitz algorithm, least squares (LS) fitting algorithm and the FFTmethod. The contrast ratio selected for the simulation is the mean value and rootmean square error of DOA estimation, and calibration residuals. Simulation resultsshows that there are little differences between the three algorithms in the case ofsmall phase error, while the FFT algorithm outperforms the LS algorithm andToeplitz algorithm in the case of large phase errors. In the case of large phase error,SNR improvement can barely improve the performance. Finally it got theconclusions that the FFT algorithm has higher robustness.In the part of super-resolution algorithm in HF hybrid sky-surface wave radar,principle of classical MUSIC and reduced-dimension MUSIC algorithm wasdescribed, and discuss the statistics deficiency of single snapshot in covariance matrix estimation. After that a simulation was performed to compare the twoalgorithms in subspace estimation accuracy under single snapshot, obtain theconclusion that the reduced-dimension MUSIC algorithm can get more accurate innoise subspace estimation. After that it proposed two possible improvements:time-domain data overlapping segment method; take the adjacent Doppler cellalgorithms. After Monte Carlo simulation analysis, and ultimately we get a betterjoint method, real-time domain data overlap segment plus reduced-dimensionMUSIC. The simulation results show that this joint approach can maintain theadvantages of both algorithms with the lowest root mean square error in the lowSNR and high stability in the high SNR.