二维桥式吊车自适应神经网络消摆控制  

Adaptive neural network anti-swing control for two-dimension overhead crane

作　　者：何熊熊[1] 王逸文 朱铮旸 陈强[1] HE Xiongxiong;WANG Yiwen;ZHU Zhengyang;CHEN Qiang(College of Information Engineering,Zhejiang University of Technology,Hangzhou 310023)

机构地区：[1]浙江工业大学信息工程学院,杭州310023

出　　处：《高技术通讯》2022年第5期454-461,共8页Chinese High Technology Letters

基　　金：国家自然科学基金面上项目(61973274,61873239)资助。

摘　　要：针对含有未建模动态和不确定参数的二维桥式吊车系统,提出一种自适应神经网络消摆控制方法。首先,基于台车位移和摆角误差设计滑模变量,使得当滑模变量收敛至零时,各误差变量均能够收敛至零点,从而保证台车精确位置控制的同时消除负载摆动。其次,设计自适应神经网络控制器,利用神经网络逼近包含未建模动态和不确定参数在内的非线性不确定性,降低对系统模型的依赖性以及避免对其线性化处理。与基于模型的吊车控制方法相比,本文所提方法不依赖系统精确模型,且兼具滑模控制的鲁棒性。最后,通过二维桥式吊车实验对比验证了所提方法的有效性。An adaptive neural network anti-swing control scheme is proposed for a two-dimensional overhead crane system with unmodeled dynamics and uncertain parameters.First of all,a sliding mode variable is designed based on the trolley displacement and swing angle error.When the sliding mode variable converges to zero,each error variable can also converge to zero,such that the accurate control of the trolley position and elimination of load swing can be guaranteed simultaneously.Secondly,an adaptive neural network controller is presented,and the nonlinear uncertainties including unmodeled dynamics and uncertain parameters are approximated by using neural networks.With the proposed controller,the dependence on the system model is reduced and the linearization of the model can also be avoied.Compared with the model-based crane control schemes,the proposed method is independent of the accurate system model,and has the robustness property of sliding mode control.Finally,the effectiveness of the proposed method is verified by a two-dimensional overhead crane experiment.

关 键 词：二维桥式吊车 自适应控制 滑模控制 神经网络 消摆控制 

分 类 号：TH215[机械工程—机械制造及自动化] TP273[自动化与计算机技术—检测技术与自动化装置]