Study Of Ocean Currents Detection With HF Ground Wave Over-the-Horizon Radar

The high frequency ground wave over-the-horizon radar can be adopted in detecting ocean state such as wind, ocean waves and ocean currents. Compared with the conventional detecting devices such as drifting buoys, submerged buoys, current meters and the modern monitoring techniques such as satellite and airborne remote sensing, the high frequency ground wave radar has the advantages of large coverage area, high precision, low investment, real time characteristics and all weather capability, therefore it is of great application values in the fields of ocean environment monitoring, resources exploiting, ocean engineering and weather forecasting. In this dissertation we conduct a detailed research on some aspects of the ocean currents detection with the high frequency ground wave over-the-horizon radar, and the main contents of this dissertation are summarized as follows.The super-resolution technique for bearing estimation of radial ocean currents is studied. The correlation of scattered signal from ocean currents is analyzed from the measured data obtained with the high frequency ground wave radar OSMAR2000. To reduce the effect of correlation on the estimation performance of the MUSIC algorithm, the super-resolution techniques based on the forward-backward smoothing technique and Toeplitzization are presented to perform the bearing estimation of ocean currents, and the principle of decorrelation with Toeplitzization is discussed with emphasis. Finally the proposed method is applied to the measured data, it is shown that by decorrelation the estimation performance can be improved remarkably.The fast realization of bearing estimation of ocean currents is investigated. First based on the theory that the Krylov subspace can be equivalent to the signal subspace under some conditions, an improved fast subspace estimation approach based on Lanczos iteration is proposed. Compared with the existing one, the computation load can be reduced effectively. To further decrease the computation complexity, the signal subspace estimation with the Multi-Stage Wiener Filter (MSWF) is studied. First a new selection method of the desired signal is presented, then an improved fast signal subspace estimation approach based on the forward recursion of MSWF is put forward, which does not need a priori knowledge of the desired signal, and compared with the existing one, the estimation performance for lower signal-to-noise ratio can be improved remarkably without extra computation complexity. To further improve the estimation performance, a fast signal subspace estimation method based on MSWF and the Rayleigh-Ritz approximation is presented, the MSWF forward recursion is utilized to obtain an orthonormal basis of the Krylov subspace, and the signal subspace can be estimated using the Rayleigh-Ritz approximation. By simulation it is clear that it can perform as well as the conventional eigenvalue decomposition does. Finally by analysis of the measured data, it is shown that the proposed methods for fast estimation of signal subspace can achieve better results.The estimation of number of sources in ocean currents detection is studied. First the influence of the forward-backward smoothing on the spreading of the noise eigenvalues is analyzed, then the Spatial Smoothing AIC (SSAIC) and Spatial Smoothing MDL (SSMDL) for estimating the number of sources is presented. By spatial smoothing the spreading of the noise eigenvalues can be reduced remarkably, hence the probability of correct detection can be increased effectively. In addition, the influence of diagonal loading on the noise eigenvalues is analyzed, and a method based on spatial smoothing and diagonal loading is put forward for estimating the number of sources, the effectiveness of which is verified by simulation.Modeling of the first-scatter from ocean currents based on distributed sources is studied. First the Doppler resolution of the scattered echoes is utilized to analyze the distribution characteristics of ocean currents, which shows that the ocean currents should be treated as distributed models. Then according to the Bragg scatter mechanism of ocean surfaces, the first-order scatter model of ocean currents based on distributed sources is proposed. With this model the distribution characteristics of ocean currents can be described effectively, and the first-order scatter of ocean currents can be simulated under different conditions, therefore it provides an effective tool to evaluate the methods for DOA estimation. Then the first-order scatter from the ocean currents with two different current profiles is simulated using this model, and by the current profile inversion the generalized MUSIC algorithm for bearing estimation of ocean currents is validated. Finally the measured data is analyzed with the generalized MUSIC algorithm, the results further demonstrate the distribution characteristics of ocean currents...