Design Of The Near-field Wideband Beamforming Algorithm

Array signal processing technique,a branch of signal processing techniques,becomes more and more important in this field as the signal processing technology improved rapidly.Therein,because of its good performance,adaptive beamforming is generally applied in many areas such as radar,astronomy,earthquake forecasting,medical diagnosis,etc.Not surprisingly,all those massive application requirements make array signal processing a hot pot for researchers.Beamforming technique implements the undistorted reception and the suppression of the interference and noise by applying sampling and weighting process to the signal that received by the array.Most of the study in this field focuses on the processing of narrowband signal in far-field environment,while the wide band signal is applied more and more widely since it has many advantages.Meanwhile,application requirements in near-field environment,such as microphone array,vehicle audio system,MRI and so forth,begin to emerge.Therefore,the study of beamforming algorithm in near-field is of great significance.In this paper,the wideband beamforming algorithm in near-field environment is studied.Also,the circular array internal beamforming performance and noise suppression problem is discussed.The main work that has been done in this paper is described below.First of all,the beamforming of near-field wideband signal inside of circular array is stated.An circular array internal hypothesis is adopted for the unform circular array.The response matrix of the array model in this paper is deduced by utilizing the near-field array response obtained by solving the wave equation along with the techniques for time-domain wideband signal processing method.Secondly,the invariance of the wideband beamforming frequency is defined by the concept of the Spatial Response Variation(SRV).On this basis,two wideband beamforming methods,based on convex optimization and least squares respectively,are introduced.The simulation results show that both the two algorithms possess good sidelobe performance and frequency invariance,while convex optimization will lead to gain loss of the expected signal.Then,the beamforming performance inside of the circular array of the LS algorithm based on SRV constraint is verified.Also an improved algorithm based on iterative variable weighted is proposed to overcome the disadvantage that it cannot meet the beamforming requirements at some specific direction.At last,the noise suppression problem in the near-field wideband signal beamforming is discussed.A LCMV wideband algorithm based on SRV constraint is proposed.We have discussed the effects of different SRV constrant range that been selected.orthogonal decomposition of weight vector are also used to improve the beamforming's sidelobe and noise performance.And the optimal solution is obtained according to deduction.The simulation results indicate that the LCMV wideband algorithm based on SRV constraint has an good frequency invariance performance as well as the suppression of sidelobe energy.Although the noise output performance is outperformed by the traditional LCMV algorithm since the new one need to take care of frequency invariance,but it's able to obtain convergent SNR output faster with a less simple taps.Also,it performs as well as the normal LCMV algorithm when the input is with high SNR.In the end of this paper,the concept of the sensors delay lines(SDL)is introduced,then we study the apply to the SDL in the near-field uniform circular array and a new array model of SDL is proposed.And conducted a simulation.