Study On Low-angle Estimation In VHF Radar

Since VHF (Very High Frequency) radar has unique advantages in anti-stealth andanti-ARM, it has attracted the world-wide attention in recent years. However, theaperture of VHF radar is limited by the long wavelength, so VHF radar has wide beam,which causes bad angular resolution. Especially when VHF radar detects or trackslow-angle target, the received signal in the mainlobe includes the direct path echo fromthe target directly and the reflection multipath echo which are reflected from the ground(sea) surface．The reflection multipath echo, consisting of specular reflection anddiffuse reflection, will reduce the measurement accuracy of elevation sharply, worsenthe performance of target localization and tracking, and result in losing the target in badcondition. At present, many scholars devoted a considerable amount of research tolow-angle estimation, and have made some breakthroughs and achievements, but thereare many technical problems have not been resolved completely. How to effectivelyimprove the low-angle estimation accuracy in VHF radar is still a problem for radarindustry.Based on previous work and combined with my research tasks at present，thisdissertation starts in the related issues of low-angle estimation in traditional array VHFradar, VHF MIMO (Multiple-Input Multiple-Output) radar and electromagneticvector-sensor array VHF radar. And this dissertation aims at the study objectives of lowcomplexity, de-correlation and high estimation accuracy in low-angle estimation. Themain content of this dissertation is summarized as follows.1. The multipath echo model in VHF radar is analyzed. The causes inducingmultipath reflection are introduced. The echo signal model of low-angle target isconstructed, including direct signal, specular reflection signal and diffuse reflectionsignal. Then, the main factors affecting the surface reflection coefficient and the diffusereflection coefficient are also analyzed, which includes signal frequency, signalpolarization, surface type, surface roughness, earth curvature, etc.2. The subspace-based low-angle estimation method in VHF radar is researched.Based on generalized MUSIC (Multiple Signal Classification) algorithm, an elevationangle and multipath fading coefficient of the target joint estimation algorithm isproposed in VHF multipath environment. In this algorithm, the virtual multipath fadingcoefficient is derived from the array covariance matrix and search angles, then asubspace is formed by this coefficient and search angles. When this subspace isorthogonal to the noise subspace, the direction-of-arrival (DOA) is achieved, and the corresponding coefficient is the real multipath fading coefficient. With the target angleunknown, this algorithm can simultaneously estimate the angle and multipath fadingcoefficient after only one time angle searching. And the surface reflection coefficientcan be achieved by the multipath fading coefficient. While overcoming multipath effectto elevation angle measurement, this method retains antenna aperture, and does notrequire a special antenna array configuration.3. The temporal-spatial sequential low-angle estimation method in VHF radar isresearched. Since most of the existing algorithms did not make full use of theinformation which can be obtained by radar and the existing temporal-spatial sequentialmaximum likelihood algorithm has large computational burden, a fast algorithm forlow-angle estimation in multipath environments for VHF radar via matrix pencil isproposed. First, by utilizing the information of range and velocity, which has beenachieved, the received data is integrated to form a single sample of the array signalvector. Then, a matrix pencil is constructed by the vector. Finally, the generalizedeigen-decomposition is employed to estimate the DOA directly. While overcomingmultipath effect, this method avoids the spectrum peak searching and reduces thecomputational burden.4. Low-angle estimation in the presence of multipath effect and mutual couplingerror in VHF radar is researched. The performance of antenna arrays is strongly affectedby mutual coupling, and ignoring its effect can seriously degrade the performance ofDOA estimation algorithms. So the signal model is constructed in the presence ofmultipath effect and mutual coupling, and an auto-calibration DOA algorithm in VHFradar is proposed. In the presented algorithm, based on the characteristic of the mutualcoupling matrix for uniform linear array, the form of the signal parameters in thereceived data covariance matrix is transformed. Without any previous mutual couplingcalibration information, a subspace-based method is applied to estimate the target angle,and simultaneously the surface reflection coefficients and mutual coupling matrix canbe achieved. Compared with many existing array calibration methods, the proposedalgorithm retains antenna aperture and avoids iterative processing under multipath andunknown mutual coupling. Furthermore, it requires neither the presence of calibrationsources nor previous calibration elements.5. Low-angle estimation in VHF MIMO radar is researched. To reduce thecomputational burden and retain estimation precision, beam-space maximum likelihoodalgorithm and beam-space generalized MUSIC algorithm are developed. In thebeam-space pre-processing method, the receiving beams and "transmitting beams" are formed at the receiving side, and the target angle is estimated in beam space. Firstly, themultipath signal model for MIMO radar is constructed with both transmitted multipathand received multipath signals being considered. Secondly, the proposed beam-spacemaximum likelihood algorithm is presented to estimate the target angle in multipathenvironment. Thirdly, the CRB (Cramer-Rao Bound) of low-angle estimation in MIMOradar is derived. Finally, to reduce the computational burden further, a modifiedgeneralized MUSIC algorithm is developed for low-angle estimation after beam-spacepre-processing. Compared with the element-space processing, the proposed method hasnearly the same estimation precision and owns lower computational burden.6. Low-angle estimation in VHF radar using vector-sensor array is researched.Firstly, the receiving signal model of vector-sensor array is constructed. Secondly, atwo-dimensional DOA and polarization estimation algorithm for coherent sources usinga linear vector-sensor array is presented. In this algorithm, a second-order statisticsmatrix is defined by the receiving data. Matrix pencil is achieved by the singular valuedecomposition of the second-order statistics matrix, and then the target elevation isestimated. Simultaneously, signal Poynting vector can be achieved by using a vectorcross-product estimator. And then target azimuth and polarization parameter can becomputed directly. If the element interval space is larger than a half-wavelength, thetarget elevation estimation can be disambiguated. The proposed algorithm owns lowcomputational complexity for avoiding angle searching, and it achieves two-dimensionangle and polarization parameter by using a linear vector-sensor array withoutparameter pairing operation. Finally, the multipath signal model of vector-sensor arrayis constructed, and the proposed algorithm is applied in low-angle estimation．Withoutangle searching, the presented method can effectively eliminate multipath effect.