Research On Robust Adaptive Beamforming And Beampattern Synthesis Algorithm

Adaptive beamforming is also known as spatial filtering and adaptive pattern synthesis technology is an important branch in array signal processing. They have been widely used in radar, sonar, seismology and radio communications and other fields. Adaptive beamforming is mainly in the retention of the desired signal and suppress interference signals and noise,while the adaptive pattern synthesis algorithm is mainly to keep the mainlobe distortion, and reduce the sidelobe level in the actual control of real-time system. Because the error is inevitable and can not be predicted, the conventional adaptive beamforming algorithm will suffer severe performance degradation. The phenomenon is called as 'signal cancellation', so the robust adaptive beamforming algorithm has been an intensive research. this paper research work mainly include:First, we discuss the array antenna signal statistical mathematical model. This paper introduces the several equivalent of beamforming optimization principles, the adaptive beamforming algorithm and adaptive beampattern synthesis technology of related parameters settings. We also introduced the basic theory of convex optimization knowledge, applied in the actual system advantages of convex optimization, gived some commonly used model and the convex optimization model used in this paper. Simple theoretical deductions and experimental analysis of several famous adaptive beamforming algorithms are also given.In the actual real-time system, the existence of the error is inevitable, and training data will contain desired signal, real interference plus noise covariance matrix is hard to obtained.This paper proposed a beamforming technique based on array steering vector estimation and interference plus noise covariance matrix reconstruction. First of all, we applied subspace theory to obtain the signal subspace, and then, we applied the array steering vector covariance matrix to obtain orthogonal matrix containing steering vector. Using the fact that real array steering vector is in the two space intersection, to obtain desired signal steering vector estimation; Then correlation coefficients between the desired signal steering vector and receiving data characteristic vector are used to estimate the array steering vector. The vector corresponding to the biggest correlation coefficients is the the vector of desired signal, and its corresponding characteristic value is the desired signal power. We set it as the average of the noise power to achieve interference and noise covariance matrix. When dispersive effects due to system bandwidth and time delays are present,a single signal source may have more than one eigenvalue associated with it. A principle eigenvalue contains most of the signal energy and one or more secondary eigenvalues contain the residual energy. We can also be calculated according to the correlation direction vector, and using integral reconstruction covariance matrix with excellent performance but large amount of computitional.Traditional data independent pattern synthesis algorithm has closed-solution design. Its poor performance cannot meet the increasing performance requirements, so the adaptive pattern synthesis algorithm is imminent. We mainly introduced the first method based on penalty function and the second-order cone programming algorithm of adaptive pattern synthesis design. Each method has advantages and disadvantages. To meet the requirements of actual system, this paper put forward a adaptive pattern synthesis algorithm based on the optimal SOCP sidelobe level, not only can get the lowest sidelobe level ,can but also make sure that the algorithm has a solution. The convex optimization software is used to help solving the optimization problem, and now the convex optimization software completely satisfy the real-time system.