Research On Wideband Constant Beamwidth Beamforming With Sensor Delay Lines

With the continuous development of wideband signal processing technology, widebandbeamforming has been widely applied to radar, communication and sonar. When the widebandsignal appeared form the non-spindle of beam, output of beam will lose some high frequencyerergy, cause the output waveform distortion. As a result, based on the wideband beamformingstructure with sensor delay lines (SDL), this paper focuses on the low computational complexitywideband constant beamwidth beamforming with sidelobe control and interference suppression,the wideband adaptive constant beamwidth beamforming based on beamspace, the design ofconstraint matrix in original wideband linearly constrained minimum variance (LCMV)beamformer and the robust wideband constant beamwidth beamforming against look directionerror. The main work and innovations obtained in this paper can be summarized as follows.1. To control sidelobe and suppress the interference, a novel wideband constant beamwidthbeamforming algorithm is proposed by using discrete spatial response variation (SRV) constraint.Firstly, SRV is computed in the discrete form. In the view of the beampattern optimization, theproposed models for controling sidelobe and suppressing the interference are provided accordingto the rule of minimizing the designed sidelobe level utmost while keeping the look directionresponse close to the unity response and the constraints of the sidelobe level, mainlobe width,SRV and so on. The constrained optimizing formula can be solved after being reformed to theform of standard second order cone programming (SOCP). The simulations illustrate that theconstant beamwidth beampattern with the overall low sidelobe and the deep null in the directionof the interference can be obtained and the computational complexity is lower.2. The design of constraint matrix in original wideband LCMV beamformer will becomplex when the signal of interest arrived from the any direction except0. To resolve thisproblem, a novel wideband LCMV algorithm is proposed with constant beamwidth property.Starting with constant beamwidth, a group of constraints imposed on the sampled frequencypoints can be replaced by one single constraint at a reference sampled frequency point. So that asimple constraint can be obtained and an expression of weight vector optimization has beenpresented at last. The simulations show the effectiveness of the proposed algorithm3. The wideband adaptive beamforming always suffers from the poor convergence speed.To alleviate this problem, a novel beamspace adaptive beamforming algorithm is proposed.Applying the beamspace technique to the SDL, wideband beamforming is processed on thebeamspace replaced the classical element space. The design utilizes the constrained optimizingby some constraints that can guarantee the frequency invariant beams. To combine each beam outputs with only one adaptive weight, the weight coefficients are reduced and adjusted by thenormalized LMS (last mean square) algorithm. The simulations illustrate that the convergencespeed of the proposed method is faster and the beampattern with constant beamwidth can beobtained.4. Look direction error can severely degrade performance of the wideband beamformer. Toimprove robustness of the wideband beamformer against look direction error, a robust widebandbeamforming algorithm is proposed by using probability-constrained optimization both in thecase of Gaussian and non-Gaussian steering vector mismatch. Based on the SDL structure, theoptimizing formula in the proposed algorithm is given by a group of probabilistic constraints andthe weight vector constraint. The computational complexity can be reduced due to the additionalconstraint to the fluctuation of the array responses within the desired frequency band. Finally, theinfluence of the number of sampling points per period, signal-to-noise ratio, and some constraintparameters on the output signal-to-interference-plus-noise ratio (SINR) and the beampattern arealso discussed in this paper. Compared with diagonal loading and the robust widebandbeamforming methods based on worst-case performance optimization, a significant improvementin frequency response consistency and a highest output SINR have been achieved in the sameexperimental conditions with the low computational complexity.