某扫雷犁电液伺服系统的HGA-RBF神经网络建模  

Modelling of the Electrohydraulic System of a Certain Mine Sweeping Plough Based on HGA-RBF Neural Network

作　　者：童仲志[1] 张媛[1] 邢宗义[1] 侯远龙[1] 贾利民[2] 

机构地区：[1]南京理工大学机械工程学院,江苏南京210014 [2]北京交通大学交通运输学院,北京100044

出　　处：《机床与液压》2009年第7期113-116,207,共5页Machine Tool & Hydraulics

基　　金：国家自然科学基金资助项目(60674001);国家重点实验室开放课题基金资助项目(SKL2008K010)

摘　　要：采用递阶遗传算法学习径向基函数神经网络的结构和参数,实现了某扫雷犁电液伺服系统的神经网络精确建模。介绍了某扫雷犁电液伺服系统的工作原理,详细论述了基于递阶遗传算法的径向基函数神经网络设计方法,即采用控制基因确定隐节点数目,采用参数基因辨识隐节点中心参数,利用最小二乘法计算连接权值。最后采用该方法对某扫雷犁电液伺服系统进行了建模研究,并与其它建模方法进行了比较,实验结果验证了该方法的有效性。A radial basis function neural network modelling method based on hierarchical genetic algorithm was proposed and applied to the hydraulic system of a certain type of mine sweeping plough. The operation principle of the hydraulic system was introduced. The neural network modelling technique using hierarchical genetic algorithm was detailed. The number of hidden units was determined by the control genes and the parameters of hidden centers were identified by the parameter genes, and the weighs were calculated using the least square method. The proposed method was applied to the electrohydraulic system, the results and comparisons with other modelling techniques clearly show its validity.

关 键 词：扫雷犁 电液伺服系统 径向基函数神经网络 递阶遗传算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TJ812.5[自动化与计算机技术—控制科学与工程]