將一種在線變結(jié)構(gòu)徑向基函數(shù)網(wǎng)絡(luò)訓(xùn)練算法應(yīng)用到內(nèi)?？刂浦?，對(duì)如何利用徑向基函數(shù)網(wǎng)絡(luò)辨識(shí)被控對(duì)象內(nèi)部模型和控制器模型進(jìn)行了較為深入的分析并給出了具體的迭代學(xué)習(xí)公式。仿真結(jié)果證明了該算法對(duì)非線性動(dòng)態(tài)過(guò)程是有效的，具有較好的自適應(yīng)性和魯棒性。
關(guān)鍵詞：RBF網(wǎng)絡(luò)；內(nèi)?？刂疲?a href='http://www.greenbey.cn/tags/神經(jīng)網(wǎng)絡(luò)/' target='_blank' class='arckwlink_none'>神經(jīng)網(wǎng)絡(luò)；變結(jié)構(gòu)
Abstract: A online varying-structure training algorithm for RBF network is used on internal model control. It is deeply analyzed that how to realize internal model and internal model controller by RBF network. Furthermore, the final iterative formula for training network is given. The simulation shows that the proposed algorithm is effective to nonlinear dynamic process, and it has excellent properties of adaptability and robustness.
Key words: RBF network；internal model control；neural network；varying-structure