基于快速终端滑模的模糊逼近系统及其应用  

A Fuzzy Approximator Based on Fast Terminal Sliding Mode and Its Application

作　　者：刘云峰[1] 缪栋[1] 刘华峰[2] 梁树晖[3] 

机构地区：[1]第二炮兵工程学院,陕西西安710025 [2]太原理工大学机电所,山西太原030024 [3]第二炮兵装备研究院,北京100085

出　　处：《上海航天》2007年第6期6-9,23,共5页Aerospace Shanghai

摘　　要：为改善模糊系统逼近的收敛速率,提出了一种基于快速终端滑模的反向传播(BP)学习算法。将快速终端滑模引入模糊逻系统的学习算法,综合终端吸引子有限时间收敛和线性系统指数收敛的性质,提高了对非线性连续函数的逼近能力。仿真结果表明:该算法能以较快速率逼近未知函数,且逼近过程稳定,较普通学习算法更为精确和有效。A back propagation learning algorithm based on fast terminal sliding mode to approximate was put forward to improve the convergence of fuzzy system in this paper. The fast terminal sliding mode was introduced into the learning algorithm of fuzzy logic system. The approximate system efficiency of unknown nonlinear continuous functions was improved with convergent property of terminal attractor in finite time and exponential convergent property of linear system. The simulation results showed that this algorithm could approximate unknown functions with rapid velocity and tranqtiilization process, which was better than conventional method in exactitude and availability.

关 键 词：模糊逻辑系统 反向传播算法 快速终端滑模 终端吸引子 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]