伺服系统的神经网络摩擦力自适应补偿研究  

Study of the Friction Compensation in Servo Systems

作　　者：张媚[1] 李秀娟[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016

出　　处：《计算机仿真》2003年第12期70-73,共4页Computer Simulation

摘　　要：在高精度伺服系统中 ,摩擦力是影响其低速性能的关键因素。该文分析了摩擦力的特性、数学模型、及其对伺服系统性能的影响 ,提出了基于RBF网络的自适应摩擦力补偿方法 ,并将其与参数线性化模型相比较。在某单轴速率 /位置转台的控制系统中的应用结果表明 ,该方法能有效地改善伺服系统的性能。In order to achieve high precision position/velocity control, friction must be appropriately compensated for. In this paper, the friction models, its impact on the servo system are analyzed, and the disadvantages of typical compensation methods are discussed firstly. Then, an adaptive friction compensation method using the RBF network is put forwarded. This compensation method is compared with the model identification method by applying them to the control of a one-axis velocity/ position simulator. It turns out that the RBF network can greatly improved the system tracking performance.

关 键 词：伺服系统 神经网络 摩擦力 自适应补偿 非线性特性 

分 类 号：TM921.541[电气工程—电力电子与电力传动] TP183[自动化与计算机技术—控制理论与控制工程]