重型机械臂电液系统的神经网络辨识与反步控制  

Neural Network Identification and Backstepping Control of Electro-hydraulic Cylinder for A Heavy-duty Manipulator

作　　者：刘汇 王恒升[1] 郭新平 孙晓宇 LIU Hui;WANG Hengsheng;GUO Xinping;SUN Xiaoyu(School of Mechanical and Electrical Engineering,Central South University,Changsha 410083,China)

机构地区：[1]中南大学机电工程学院,长沙410083

出　　处：《机械科学与技术》2023年第11期1767-1777,共11页Mechanical Science and Technology for Aerospace Engineering

摘　　要：针对重型机械臂的电液驱动系统因非线性、参数时变等因素引起的控制精度下降问题,提出了一种基于RBF神经网络辨识动态负载的反步控制策略。以某锚杆钻车重型机械臂的电液系统为例,建立了系统的数学模型,将其分解为系统内部状态反馈、控制器驱动及外部负载驱动这3个动力学部分。考虑电液系统内部参数变化的缓慢性,通过离线辨识的方法,得到系统内部状态反馈中的标称模型参数。控制器的设计采用反步法,其输出计算需要对外部负载进行辨识,对此采用RBF神经网络进行动态负载辨识,辨识与控制的动态过程及设计原则依据Lyapunov稳定性原理。仿真与实验结果表明:所设计的控制算法有效提高了机械臂的位置跟踪精度,具有响应速度快、轨迹误差小的特点;控制器输出的控制量也相对较小和平滑。For the problem of control accuracy drop of the electro-hydraulic system used widely in heavy-duty manipulators because of the factors such as nonlinearity and time-varying parameters,a backstepping control method is proposed based on the introducing of Radical Basis Function(RBF)neural network to identify the dynamic load.Taking the hydraulic-driving manipulator of an anchor drilling rig as an example,the mathematical model of the electro-hydraulic system is established.The model consists of three dynamic parts,which are inner state feedback of the system,control driving and outside load driving.The nominal model parameters in the internal state feedback of the system are identified through off-line test due to the relatively slow variation of the parameters.The design of the controller follows the procedure of backstepping method,and the output calculation of the controller needs the estimation of the dynamic load,for which a RBF neural network is adapted.The principle for the update of RBF neural network and the design of controller is based on Lyapunov theory.Many simulations and experiments show that the proposed control method effectively improves the position tracking accuracy of the manipulator,with rapid response and less trajectory error,and relatively smooth and small control output.

关 键 词：重型机械臂 电液驱动系统 RBF神经网络 动态负载辨识 反步控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]