模拟电路行为级建模的SGWD逼近算法  

Behavioral Modeling of Analog Circuit via SGWD Approximation

作　　者：陶俊[1] 曾璇[1] 蔡伟 

机构地区：[1]上海复旦大学微电子系专用集成电路与系统国家重点实验室,上海201203 [2]Department of Mathematics University of North Carolina Charlotte

出　　处：《计算机辅助设计与图形学学报》2007年第5期622-627,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(90307017;60676018);国家"九七三"重点基础研究发展规划项目(2005CB321701);教育部跨世纪优秀人才培养计划基金;教育部高等学校博士学科点专项科研基金(20050246082).

摘　　要：提出采用一种规范化的香农小波(distributed approximation function,SGWD)逼近算法建立模拟电路单元模块的行为级模型.为降低逼近的边界误差,同时提出了周期性展开和偶对称映射2种预处理算法.周期性展开算法需要对电路模块的原始I/O函数做一定变形,构造出周期函数后再进行逼近;偶对称映射算法则是将I/O函数在边界点处进行偶对称翻转,然后采用SGWD逼近算法建模.与传统的建模方法,如多项式逼近等算法相比,该方法结合了边界误差降低技术的SGWD逼近算法具有更低的计算复杂度,并能达到更高的建模精度.In this paper, Shannon-Gabor wavelet distributed approximation functions （SGWD）, which combine Shannon sampling with a Gabor-distributed approximation functional （Gabor-DAF） window function, are used to construct the behavioral models of analog circuit blocks. Two techniques, periodic extension of the approximated curve and even-symmetrical mirroring of the original transfer function, are proposed to reduce the approximation boundary errors. In the first technique, the approximated function is modified to make the two boundary values equal, and then the modified curve is expanded into a periodic function. In the second algorithm, the approximated curve is expanded by even symmetrical mirroring at the two boundary points. Compared with the classical polynomial approximation method, the proposed methods have less computational complexity and can reach higher accuracy. Some numerical examples are also given to demonstrate the merits of these proposed algorithms.

关 键 词：SGWD逼近 模拟电路行为级建模 边界误差 

分 类 号：TN47[电子电信—微电子学与固体电子学]