Research On Wideband Frequency Invariant Beamforming Algorithm

In recent years,the amount of broadband signals due to target information carrying less correlation,is conducive to the estimation and target extraction and target detection parameters of many points,which makes the broadband signal processing has become an important research direction in the field of signal processing.The broadband beamforming technology has became the array signal processing research hotspot in the field of which are widely used in communication,radar and sonar etc.The technology of frequency invariant beamforming is essentially how to obtain the weighting coefficient and filter coefficient of the sensor array.Based on the analysis and summary of the research both at home and abroad,aiming at the wideband signal processing problem,based on the detailed discussion of broadband beamforming,the broadband frequency invariant beamforming technology is studied in depth.1.Firstly,introduce several broadband beamforming algorithms.Then,a Wiener post filter based least squares beamforming reverberation suppression algorithm is proposed based on broadband beamforming algorithm.The algorithm of speech signal after the reverberation estimation Wiener post filter gain for the direct part and the part of speech signal to reverberation;strong noise in low frequency coherence,frequency division processing forming algorithm is using the least square beam,and finally solve the optimal weights.2.In order to achieve the purpose of signal transmission without distortion,the array response is not affected by the frequency change of broadband FIB,including linear constrained minimum variance frequency invariant beamforming algorithm and least squares frequency invariant beamforming algorithm.Based on the theoretical analysis of the above 2 algorithms of mathematics the formula and algorithm simulation.Focuses on the FIR filter frequency invariant beamforming algorithm based on the algorithm using FIR filter array structure and symmetry properties of uniform line array beamforming,determine the wide range of frequency algorithm and obtain approximate constant beam array constant beamwidth in response to fluctuations;due to the problem of the cut-off frequency between adjacent beam width,we design the reference beam and the optimal weights obtained by least square method;finally the optimal weights are normalized to obtain uniform gain.3.Typical transient interference,such as knocking on door and keystroke,which is short time,frequency domain widely distributed,and with sudden characteristics.Traditional speech enhancement technology is not enough to make speech enhancement by short time spectrum estimation method.In this paper,LCMV-FIB speech enhancement algorithm based on logarithmic spectrum estimation is proposed,which is used to suppress transient and pseudo stationary background noise.The algorithm consists of two stages:first,invariant beamforming algorithm without distortion of the received speech signal and spatial filtering of the speech signal using linear constrained minimum variance frequency;second stages,using the optimized modified log spectral estimation algorithm to suppress the residual noise and the ideal signal from the same direction,to achieve the purpose of speech enhancement.Experimental simulation waveforms and speech quality evaluation indicators show that the algorithm has good robustness,and can effectively suppress instantaneous interference and background noise,and achieve the purpose of speech enhancement.