论文部分内容阅读
为了解决数字信道化中窄过渡带滤波器组计算复杂度较高的问题,提出了一种采用FRM的窄过度带奇型非最大抽取滤波器组高效结构。在设计该结构的原型滤波器时,首先通过对原型滤波器进行插值得到FRM上支路的滤波器,然后将信号延时与插值后的滤波器相减得到位于下支路的互补滤波器,再利用屏蔽滤波器对插值产生的多余镜像进行屏蔽,从而最终得到我们所设计的FRM滤波器。对该改进结构进行了MATLAB仿真,仿真结果验证了该改进结构的正确性,并且采用该改进方法设计的高效结构相比多相滤波器组结构节省了79%的乘法器资源。由于该改进结构中将抽取提前,降低了系统采样率的限制条件,可以直接应用于高速采样系统中。
In order to solve the problem of high computational complexity of narrow-transition band filter banks in digital channelization, a novel high-efficiency narrow-band non-maximal extraction filter bank with FRM is proposed. In designing the prototype filter of this structure, the filter of the upper branch of the FRM is first interpolated by the prototype filter, and then the signal delay and the interpolated filter are subtracted to obtain the complementary filter located in the lower branch. Then the mask filter is used to shield the extra image generated by the interpolation so as to finally obtain the FRM filter we designed. The improved structure is simulated by MATLAB. The simulation results verify the correctness of the improved structure. The high efficiency structure designed by the improved method saves 79% of the multiplier resources compared with the polyphase filter bank. As the improved structure will be extracted in advance, reducing the system sampling rate constraints, can be directly applied to high-speed sampling system.