基于在线聚类的机器人自适应模糊建模

针对机器人这种具有时变、强耦合和不确定性的复杂非线性被控对象,提出一种基于在线聚类的模糊自适应方法用于机器人系统建模。建模过程中采用在线聚类算法辨识机器人T-S模型的前提参数,采用递推最小二乘算法(RLSE)辨识结论参数,根据过程中新的数据信息,模糊规则可以自动增加、修改和删除,实现了模型结构和参数的在线辨识和更新,而且该方法具有模型结构简单、建模速度快、精度高等优点。最后通过二自由度机器人仿真研究证明了该方法的有效性和优越性。

Adaptive fuzzy modeling of robot manipulators based on on-line clustering

A fuzzy modeling approach based on on-line clustering is presented for a family of complex systems with severe nonlinearity such as of robot manipulator.This fuzzy model(Takagi-Sugeno fuzzy system) is identified by on-line clustering and recursive least square estimation(RLSE).Using this method,the rules can be added,modified and deleted with the new data information automatically,achieving the model structure and parameters' on-line identification and update rapidly and accurately.In addition,only the local model needs to be identified,which makes the model simple and accurate.The simulation results of two-DOF robot demonstrate the effectiveness and advantages of this approach.