机械臂的改进区间二型模糊神经网络控制

针对具有不确定性模型参数的双关节机械臂系统，提出基于改进区间二型模糊神经网络逼近器的自适应反演控制算法.相比于一型模糊系统，区间二型模糊系统由于自身的区间前件和隶属函数，更有效地处理高度非线性系统.然而现有的二型模糊寻找上下输出的交叉点过程中KM迭代算法计算量大、耗时高，使得传统的二型模糊系统不适用于实际控制应用.利用自适应调节因子代替KM迭代算法，在上输出和下输出建立起自适应连接，所采用的改进区间二型模糊神经网络逼近器有效解决双关节机械臂系统中不确定性参数的问题.通过李雅普诺夫方法证明了所有信号的有界性以及闭环系统的稳定性.最后仿真结果表明，基于改进区间二型模糊神经网络逼近器的自适应反演控制器可实现快速响应、更短的稳定时间和更高的跟踪精度.

Improved Interval Type2 Fuzzy Neural Networks Control of Manipulator

In this paper, an adaptive backstepping control with improved interval type2 fuzzy neural networks approximator is proposed for double link manipulator system. Compared with type1 fuzzy system, interval type2 fuzzy system can obtain better performance for highly complex nonlinear systems because of its uncertainties in the antecedent and membership functions. However, KM algorithm in the process of finding the cross points between upper output and lower output in type2 fuzzy system leads to heavy computational burden and time consumption which makes it difficult to apply in practical engineering. In this paper, instead of KM algorithm, an adaptive factor is used to build an adaptive ponderation between upper output and lower output. The improved interval type2 fuzzy neural networks approximator is applied to handle the uncertain parameters in double link manipulator system. The ultimately boundedness of all signals and the stability of the closed loop system can be mathematically proved by the Lyapunov stability analysis. The simulation results demonstrate the proposed adaptive backstepping control with improved interval type2 fuzzy neural networks approximator obtains transient response, short stabilizing time and high approximation accuracy.