基于离散泰勒级数的新型联想记忆系统  

Associative memory system based on discrete taylor series

作　　者：王俊松[1] 徐宁寿[2] 崔世钢[1] 邴志刚[1] 

机构地区：[1]天津职业技术师范学院自动化系,天津300022 [2]北京工业大学自动化系,北京100080

出　　处：《天津工业大学学报》2003年第4期56-58,共3页Journal of Tiangong University

基　　金：天津市自然科学基金资助项目(013602811)

摘　　要：基于离散泰勒级数提出一种对任意阶多维函数可实现无差逼近的新型联想记忆系统,详细讨论了该系统的插值算法、训练规则及寻址机制.新型联想记忆系统相对于CMAC具有逼近精度更高、占用存储单元较少、学习速度较快等优点.该系统在信号处理、模式识别、过程建模及高精度实时智能控制等领域具有广泛的应用价值.仿真实例表明了系统的可行性与有效性.This paper proposes a novel highorder associative memory system based on the Discrete Taylor Series(DTSAMS),which includes the interpolation algorithm and training algorithm. It is capable of implementing errorfree approximations to multivariable polynomial functions of arbitrary order. The advantages are:highprecision of learning and much smaller memory requirement. The novel associative memory system has great potential in the application areas of signal processing, pattern recognition, process modeling and implementation of highprecision realtime intelligent controllers. The simulation results verify that DTSAMS is feasible and efficient.

关 键 词：离散泰勒级数 联想记忆系统 插值算法 训练规则 

分 类 号：TN052[电子电信—物理电子学] TP301.4[自动化与计算机技术—计算机系统结构]