基于SVM的上肢康复机械臂轨迹跟踪预测控制

针对上肢康复训练机械臂具有强耦合、非线性和时变的特点，设计了基于SVM（支持向量机）的轨迹跟踪预测控制器。采集机械臂系统的输入和输出数据，通过SVM辨识得到广义逆系统，与原系统串联实现解耦。对解耦后的系统，采用SVM辨识预测模型和PSO优化滚动控制序列的预测函数控制方法，并从其内模结构分析了系统的稳定性和鲁棒性。实验结果表明，该方法能够平稳高精度地实现轨迹跟踪。

Trajectory Tracking Predictive Control of Upper Limb Rehabilitation

Due to the strong coupling , nonlinearity and time variance of upper limb rehabilitation training robot arm , the predictive controller is designed with the core of SVM. And by collecting input and output data of the manipulator, the generalized inverse system is identified with the help of SVM, which decouples with the original system. The decoupled systems adopt PFC control based on SVM identification of prediction model and predictive functional control sequence of PSO rolling optimization. Stability and robustness is analysed in term of internal model structure, experiment results show that the control method can achieve stable and higher precision trajectory tracking.