面向子带分解的非均匀抽取均衡算法

Non-uniform Extraction Equalization Algorithm for Sub-band Decomposition

宽带数字阵列雷达通道失配严重的情况下,带宽分割法相比逆傅里叶变换法均衡效果更好,但其在不完全分解时重建误差较大且占用硬件资源过多;针对上述问题,提出一种延长带宽非均匀抽取均衡算法;该方法设定需要划分的子带数目,根据阻带截止频率的位置合并不参与均衡的子带;依据待均衡子带的频率特性对其非均匀抽取,并进行子带内均衡;均衡后各子带经过插值和合并操作,完成整个均衡过程;实验结果表明,在均衡子带个数和重建误差相等的情况下,相比带宽分割法,提出方法的幅度改善因子和相位改善因子分别提高5.39 dB和7.48 dB,降低了运算复杂度和硬件资源占用率。

In the case of serious channel mismatch of wideband digital array radars,a band-partition method has a better equalization effect than an inverse Fourier transform method.However,with a decomposition is incomplete,its reconstruction error is significant,and the required hardware resources are excessive.Aiming at the above problems,an extended bandwidth non-uniform extraction equalization algorithm is proposed.This method sets the number of sub-bands to be divided,the position of stopband cutoff frequency is used to merge the sub-bands that do not participate in equalization.The non-uniform decimation is performed according to the frequency characteristics of the sub-bands to be equalized,and the equalization is performed within the sub-bands.After the equalization,each sub-band is interpolated and merged to complete the equalization process.Experimental results show that compared with the band-partition method,the proposed method increases the amplitude improvement factor and phase improvement factor by 5.39 dB and 7.48 dB,respectively,and reduces the computing complexity and occupation rate of hardware resources under the same equalized sub-bands and reconstruction error.