Research On Efficient Synthesis Method Of Wideband Frequency-Invariant Beamforming

Beamforming is one of the key technologies in array signal processing,which is widely used in wireless communication,sonar,radar and other fields.With the increasing complexity of the signal processing environment and the rapid development of broadband applications,beamforming is required to have a larger frequency bandwidth to meet practical applications.In wideband beamforming,sensor arrays usually need to receive wideband signals without any distortion,which requires the frequency-invariant beamforming(FIB)technique.Therefore,FIB has become an important research direction in the field of array signal processing recently.This article has conducted an in-depth study on efficient synthesis method of wideband FIB.The main work includes:1)The filter-and-sum array model of the FI beamformer is presented;two classic FIB techniques are studied,including the convex optimization-based FIB and the Fourier transform-based FIB,and the detailed derivation process is given;The synthesis examples of these two classic methods are given,and the synthesis results intuitively reflect the advantages and disadvantages of both methods.2)The problems existing in the classical Fourier transform-based FIB is discussed in detail,including the problems caused by matching reference pattern and the inherent problems of this method.An improved solution is provided for the former,the improved method can synthesize FI pattern with space-frequency notching.In addition,the main steps of this method are implemented by fast Fourier transform(FFT).Two synthesis examples are conducted to verify the effectiveness of the proposed method.The comparison with the convex optimization-based FIB shows the great efficiency of the improved Fourier transform-based FIB.3)The construction of the spatiotemporal Fourier transform(STFT)relationship between the wideband pattern and the filter coefficients is provided,and then an iterative STFT-based FIB is proposed,which uses two-dimensional FFT to accelerate the evaluation.The modifications used during the iteration are designed for synthesizing the focused FI pattern,including the mainlobe FI correction and the wideband sidelobe correction.A set of synthesis examples obtained by the iterative STFT method are demonstrated.The comparison with the classical Fourier transformbased FIB and convex optimization-based FIB highlights the advantages of the proposed method.4)In the procedure of the iterative STFT method,the modifications are replaced by an improved mainlobe FI correction and a new pattern shaping correction,so that the method has better synthesis efficiency and can be used for shaped FI pattern synthesis.The alternating projection method is extended to the shaped FI pattern synthesis and compared with the iterative STFT method.The comparative results validate the excellent performance of the iterative STFT method in terms of both computational and storage efficiency.