基于LS_SVM的下行波束盲校正算法

Blind Calibration Algorithm for Down Wave-beam Based on Least Squares Support Vector Machine

在智能天线系统中,由于各阵元之间存在着耦合和激励误差,在对下行波束进行赋形时,这些误差会使产生波束指向严重偏离期望的方向,造成天线增益下降。针对这一问题,提出了一种基于最小二乘支持矢量机(LS_SVM)的下行波束盲校正算法:该算法首先利用发射的训练序列对支持矢量机(SVM)进行训练,然后利用所得的权值补偿耦合和激励误差,使发射波束的指向符合期望的方向,最后采用自适应波束综合算法产生主波束宽度以及副瓣电平符合期望值的方向图。对该方法进行了仿真验证,仿真结果表明,该算法简单易行,具有实时处理的能力。

The electromagnetic field coupling and exciting errors existed in these antenna-array elements of the intelligent antenna system would result in serious direction errors of down-wave-beam between the in-factdirection and desired direction and in the decline of antenna gain when the down wave-beam is shaped.For resolvingthese problems,a novel blind calibration algorithm for down wave-beam based on LS-SVM （least squares supportvector machine） is proposed.The antenna array first transmits a serial train sequence to train the LS-SVM,andthus the desired weight value and coefficients are obtained,those weight value and coefficients are used tocompensate these array errors and make the transmitting direction in conformity with the desired direction.Finallythe adaptive pattern synthesis algorithm is adopted to produce a certain-width main beam and lower-level sidebeam.The proposed algorithm is simulated with Matlab simulink software,and the simulation result shows thatthe LS-SVM-based algorithm is simple and feasible,and is of the real-time processing ability.