Robust Broadband Beamforming Techniques

The beam former of array signal processing is a space filter. The beam former canenhance the signal in desirable direction and suppress the interference in undesirabledirection. In the early years, beam former is mainly designed in narrow band, andapplied in radar, communication etc. However, in signal processing, many broadbandsignals will be processed, for instance: earthquake signal, sound signal or sonar signal.In addition, broadband signal is used in radar and communication more and more. Innarrowband beamforming, the coefficient is correlative with frequency. If thebroadband signal is dealt with the narrow beam former, the serious disorder of thesignal will be occurred. This paper mainly studies the robust broadband beamforming. The content can beoutlined as follows: the introduction of broadband beamforming, the study of a robustbroadband generalized side lobecanceller (GSC), the discussion of an integralconstraint beamforming based on subband processing. The broadband beamforming can be designed in time domain or frequencydomain. The GSC is the most popular method in time domain. The GSC can retainhigh performance in interference-reduction under the condition of less array elements.But when the steering vector errors are in existence, the target-signal cancellation isoccurred. The steering vector errors are inevitable in actual arrays. And target signalcancellation is a serious problem. Many signal processing techniques have beenproposed to avoid signal cancellation. But they avoid the cancellation with the cost ofdegradation in interference-reduction performance. The paper studies a robust adaptivebeamformer to avoid the cancellation. In the design of the adaptive beamforming,linear constrained minimum variance is the commonly used method. This paper studiesan integral constraint beamformer to overcome the shortcoming in multi-pointconstraint beamformer that requires a large number of the degree of freedom in theweight vector. This is a two-objective optimization problem. To solve the problem, twomethods are presented, which are the weighted method and the eigenspace method.The results of simulation show that the beamformer can reduce the use of degree offreedom and can improve the performance.