【摘要】 对一种柔性关节微操作机器人系统提出了多输入多输出直接自适应模糊广义预测控制方法,此方法先基于机器人理论模型设计出广义预测控制器,再构造直接自适应模糊控制器逼近广义预测控制器,并用机器人视觉误差信息对控制器参数和广义误差向量估计值中的未知向量进行自适应调整,以增强对建模误差的鲁棒性,并证明了所设计的控制器可使微操作机器人跟踪时变参考轨迹时的广义误差估计值收敛到原点的小邻域内,以达到控制要求,仿真结果验证了此方法的有效性.更多还原

【Abstract】 A multi-input multi-output direct adaptive fuzzy generalized predictive control method for a micromanipulator with flexible joints is presented.First,a generalized predictive controller is designed based on the manipulator’s theoretical model.Then,in order to increase the robustness for modeling errors,a direct adaptive fuzzy controller is constructed to approximate the generalized predictive controller,and the fuzzy controller parameters as well as the unknown vectors in the generalized error estimation vector are adjusted adaptively by the use of vision error information.It is proved that,when the micromanipulator tracks a time—varying reference trajectory,the proposed controller forces the norm of the generalized error estimation vector to converge to a small neighborhood of the origin,thus the control requirements can be fulfilled.Simulation results demonstrate the effectiveness of this method.更多还原