摘要
针对并联机器人数学模型不完全确知并包含外部扰动的非线性多变量系统,提出一种基于模糊神经网络运算法则(FNNA)的自适应控制策略。将各个支链的模糊规则通过神经网络进行在线训练并得出模糊规则的权重并将此运用于在线辨识非线性自适应控制系统的未知动态,有效抑制了系统的数学模型不精确所产生的误差及外部扰动。仿真结果表明该控制方法明显提高了控制系统的轨迹跟踪性能,并对外部干扰及系统的非线性具有很强的鲁棒性。
As the model of parallel manipulators is nonlinear multi-variables system with structure uncertainties and external disturbances, this paper considers an adaptive control strategy with Fuzzy Neural Network Algorithm （FNNA）. The fuzzier rules of each limb can be trained on-line by neural network algorithm and weights of rules can be got. These results can be used to the unknown dynamic of nonlinear adaptive control system and effectual restrain errors which caused by the model uncertainties and internal disturbs. Simulations indicate that this method can enhance obviously the capability of track training of control system and have strong robustness with external disturbs and the nonlinear characteristic of mechanical system.
出处
《自动化技术与应用》
2010年第1期4-7,共4页
Techniques of Automation and Applications
基金
国家自然科学基金资助项目(编号50965007)
