基于神经网络的机械臂任务空间滑模同步控制

Sliding mode synchronous control for multiple robotic manipulator systems based on neural network

针对存在不确定性的多机械臂系统,运用RBF神经网络,设计了一种新的滑模同步控制器,解决了多机械臂同步运动问题。根据无向图理论,定义机械臂之间的同步误差和交叉耦合误差。使用自适应律在线更新RBF神经网络权值,逼近并补偿机械臂的运动学不确定性和动力学不确定性。根据Lyapunov方法进行了稳定性分析。最后通过仿真验证了同步控制器的稳定性和有效性。

In view of the uncertainty of the multiple robotic manipulator systems,a new sliding mode synchronous controller was designed based on RBF neural network to solve the problem of synchronous motion of multiple robotic manipulator systems.Synchronization error and cross coupling error among robotic manipulators were defined in terms of undirected graph theory.By updating the weights of RBF neural network online with the adaptive law,the kinematic and dynamic uncertainties of the manipulators were approximated and compensated for effectiveness.Afterwards,the stability of the systems was analyzed by using Lyapunov method.Finally,the stability and effectiveness of the proposed synchronous controller was validated by simulation.