一种设计M通道双正交过采样图滤波器组的新算法

发布时间：2018-08-09 10:59

【摘要】：针对现有的M通道过采样图滤波器组整体性能较差的问题,该文提出一种过采样图滤波器组设计的新算法。在新算法中,分两步来设计图滤波器组。首先,从频谱特性方面考虑来设计分析滤波器,以分析滤波器的通带波纹和阻带能量为目标函数,以3 d B约束为约束条件,通过半正定规划求解出频谱选择性较好的分析滤波器;然后,从完全重构特性方面考虑来设计综合滤波器,以综合滤波器的阻带能量为目标函数,以完全重构条件为约束函数。上述两个约束优化问题都是半正定规划问题,都可有效地求解。新算法综合考虑了滤波器组的重构特性和频率特性,因此可以设计得到整体性能良好的M通道双正交过采样的图滤波器组。仿真对比表明,与已有的设计算法相比,新算法设计所得的图滤波器组具备更小的重构误差。
[Abstract]:In view of the poor overall performance of the existing M-channel over-sampling graph filter banks, a new algorithm for the design of over-sampling graph filter banks is proposed in this paper. In the new algorithm, the filter banks are designed in two steps. Firstly, an analytical filter is designed from the aspect of spectrum characteristics. The objective function is to analyze the passband ripple and stopband energy of the filter, and the constraint of 3 dB is taken as the constraint condition. An analytical filter with better spectrum selectivity is obtained by using positive semidefinite programming, and then the synthesis filter is designed from the aspect of complete reconstruction characteristics. The stopband energy of the synthesis filter is taken as the objective function and the complete reconstruction condition is taken as the constraint function. The above two constrained optimization problems are both positive semidefinite programming problems and can be effectively solved. The new algorithm takes into account the reconstruction and frequency characteristics of the filter banks, so we can design the M-channel biorthogonal over-sampling graph filter banks with good overall performance. The simulation results show that the new algorithm has a smaller reconstruction error than the existing algorithm.
【作者单位】：桂林电子科技大学信息与通信学院;
【基金】：国家自然科学基金(61261032,61371186) 桂林电子科技大学研究生教育创新计划项目(2017YJCX21)~~
【分类号】：TN713

【相似文献】