空间机械臂刚柔耦合动力学模拟及小波基模糊神经网络控制

The Rigid-Flexible Coupling Dynamics Simulation and Wavelet Based Fuzzy Neural Network Control for Space Manipulator

研究了关节和杆件双重柔性影响下空间机械臂系统的动力学模拟,运动控制算法设计和臂杆、关节双重柔性振动的主动抑制问题。利用线动量、角动量守恒关系,并基于拉格朗日方程、线性扭转弹簧及假设模态法建立系统动力学模型。为解决传统奇异摄动法应用受关节柔性限制问题,引入了一种关节柔性补偿控制器,以适当提高系统关节的等效刚度;并基于奇异摄动理论,将系统分解成独立时间尺度的柔性臂子系统和电机力矩动力子系统,借助小波基模糊神经网络的优良特性,将轨迹跟踪的角度误差及角速度误差输入到网络中,通过一阶梯度寻优算法来优化整个网络,降低柔性振动的影响,以达到高精度控制的要求的。计算机数值仿真对比试验证实了该方法的可靠性和有效性。

The dynamics simulation of flexible joints and flexible arm space robot system,motion control algorithm design and hierarchical points order active inhibition problem of arm and joints double flexible vibration that all under the situation of parameter uncertain were discussed in this paper.With the conservation relationship of linear and angular momentum,a system dynamics model was established by Lagrange equations,linear torsion spring and hypothesis modal method.To solve the problem that the application of traditional singular perturbation approach was limited by joint flexibility,a joint flexibility compensation controller was introduced,which could properly enhance the equivalent stiffness of joints.Then,based on singular perturbation theory,the whole system was resolved into flexible arm subsystem and motor moment power subsystem on the basis of joint flexible compensation controller and singular perturbation technology.Taking advantage of the excellent characters of the wavelet fuzzy neural network,the authors input the trajectory tracking angle error and angular velocity error to the network,to optimize the whole network by a gradient optimization algorithm,so as to solve the influence of flexible vibration,thus achieve high precision control requirements.Computer numerical simulation comparison experiment testified the reliability and availability of this scheme.