摘要
给出了基于强跟踪滤波器的Jordan神经网络训练方法,该方法是一种新的学习算法。Jordan网络可以表示输出的动态特性,改进后还可以反映状态特性,更适于动态系统辨识。强跟踪滤波具有鲁棒性好、收敛快等优点,将两者结合可以得到很好的辨识效果。最后,通过仿真实例验证该方法的有效性。
A new kind of training method of Jordan neural network is given, which is based on strongtracking filter. The Jordan network is able to express the dynamic characteristic, theimproved network can also reflect state characteristic, and is suitable to dynamic systemidentification. Strong tracking filter has the strongpoint of the robustness is better and theconvergence is quick, and when they are integrated, the better identification effect is gained.At last, a model identification example is given to show the effectiveness of the method.
出处
《制造业自动化》
北大核心
2005年第3期16-18,共3页
Manufacturing Automation
基金
国家自然科学基金资助项目(60474019)。
关键词
Jordan网络
强跟踪滤波器
系统辨识
学习算法
Jordan neural network
strong tracking filter
system identification
studying arithmetic
