Research On Key Technologies Of Adaptive Digital Beamforming

Adaptive digital beamforming (ADBF) technique is one of the important branches in array signal processing which has been applied to many applications such as radar, sonar, wireless communication and radio astronomy. After several decades of developments, the fundamental theories and algorithms of ADBF technique has been presented, but the investigations on ADBF technique under complex background are of crucial importance. This dissertation investigates on ADBF algorithm in impulse noise, Adaptive Pattern Control (APC) algorithm, robust ADBF algorithm, and Compressive Sensing (CS) based ADBF algorithm. The main works of this dissertation are accomplished as follows:1. Research on ADBF algorithms in impulse noiseNumerous experimental results show that many types of noises in actual environments have impulse characteristics, in which the second-order-statistics of the input signal does not exist, so conventional ADBF algorithm can not be applied to impulse noise environments directly. Aiming at solving this problem, the normalized generalized sidelobe canceller is presented firstly. To improve the performance of the beamformer amid impulsive noise, the proposed algorithm filters the input signal which was infinity-norm snapshot-normalized by wiener filter. Secondly, to solve the problem of high sidelobe in the adaptive beamformer by using of Sample Covariance Matrix (SMI) algorithm, the normalized linearly constrained eigencanceler is given. By putting the weighted vector into noise subspace by performing Eigen Decomposition (ED) to covariance matrix of the input data which was infinity-norm snapshot-normalized, the presented algorithm improves the performance of adaptive beamformer, which is computed by using of SMI algorithm.2. Research on APC algorithmsADBF can suppress the interference, but have higher sidelobe than quiescent beamformer. To solve this problem, an APC algorithm based on transformation matrix is proposed. The proposed algorithm forms a transformation matrix to suppress the interference, then the beam is formed using the transformation matrix at the maximum Signal to Noise Ratio (SNR) criterion, so the pattern can be identified equal to the quiescent pattern finally. Secondly, by amending constraint matrix and constraint response vector of quiescent weighting vector to restrain interference and make the overall beamformer sidelobe response equal the desired quiescent response, the eigenvector-based linearly constrained minimum variance adaptive pattern control algorithm is presented. Finally, an adaptive pattern control algorithm based on Singular Value Decomposition (SVD) is proposed. The proposed algorithm can obtain the interference subspace and its orthogonal subspace by performing SVD to the receiving data matrix, then amends constraint matrix and constraint response vector of quiescent weighting vector by the interference subspace. It can restrain interference and make the overall beamformer sidelobe response equal to the desired quiescent response. Also, it is shown that the computational complexity could be reduced significantly.3. Research on robust ADBF algorithmsADBF algorithms can suppress the interference, but there are still challenges for the deviation of mainlobe and high sidelobe in the adaptive beamformer with finite snapshots and Array Steering Vector (ASV) error. In order to solve this problem, a robust beamforming algorithm is presented firstly. The presented algorithm amends the exponent of covariance matrix of input signal in the Linearly Constrained Minimum Variance (LCMV) adaptive beamformer to improve the performance of beamformer with finite snapshots. Secondly, an ADBF algorithm based on adaptive transformation matrix is also given. Based on the null information, the null preprocessing technique is applied to reform new adaptive transformation matrix for beamforming. The algorithm can keep system freedom with computational complexity reduced significantly. Thirdly, a robust adaptive beamforming algorithm based on SVD is proposed, the beam can be formed effective with finite snapshots and the proposed algorithm can suppress the interference with only one snapshot in case of one interference. Fourthly, a robust adaptive beamforming algorithm based on spectral analysis is provided. The proposed algorithm uses the spectral analysis technology to get the main beam width based on ASV error, and then uses the Second-Order Cone Programming (SOCP) technology to form a flat response in the main beam and suppress the interference in the sidelobe. Fifthly, to solve the degradation of robustness against array steering vector error with finite snapshot in the adaptive beamformer, a novel robust adaptive beamforming based on steering vector estimation and norm constraint is proposed. The presented algorithm chooses the signal subspace according to the uncertainty of the array steering vector, and then estimates the real steering vector by projecting the ideal steering vector onto the signal subspace, the SOCP technique with norm constraint on the array weight vector is lastly used to enhance the robustness of the new beamforming with finite snapshots. Finally, a beamforming adapting to ASV error with finite snapshot named robust adaptive beamforming based on worst-case and norm constraint is proposed. In order to keep the desired signal with the maximum gain, the proposed algorithm forms a flat response in the main beam, and then improves the robustness of the beamforming with norm constraint in the condition of finite snapshot.4. Research on CS based ADBF algorithmsCS has been widely appreciated for its superiority since its birth. In the present work we combine the CS with array signal processing techniques and propose ADBF algorithm based on CS. Firstly, a novel array module based on CS is presented, and then the sparity of the array signal is analyzed, at last the CS based single-snapshot adaptive pattern coltrol algorithm is derived. The presented algorithm performs better than traditional adaptive beamforming algorithms for suppression of interferences with any coherence in the condition of only one snapshot. Secondly, a novel single-port robust adaptive beamforming based on CS is presented. A single-port array system is given, and then the array signal module based on CS is build, in which the sensing matrix satisfies Restricted Isometry Property (RIP). The presented algorithm can suppress interferences with any coherence and enhance the robustness with only one single-port of radio frequency and finite snapshots.