在三级倒立摆(TIP)系统中,应用神经网络与模糊控制相结合的自适应神经模糊推理系统(adaptiveneuralfuzzyinferencesystem),根据样本数据调整隶属函数和控制规则参数,使得训练后ANFIS控制器很好地模拟期望的输入输出数据.仿真结果表明所设计的ANFIS控制器对三级倒立摆系统的稳定控制是可行的.与LQR控制相比,基于ANFIS控制的倒立摆系统具有良好的动态性能和抗干扰性能.
In the triple inverted pendulum(TIP) system, adaptive neural-fuzzy inference system(ANFIS) approach is atilized to combine fuzzy logic with Neural-Network, according to the input/output data, so that ANFIS automatically adjusts fuzzy rules and membership functions based on state synthesis to fit sampling data. The simulation results show that the designed ANFIS controller is feasible. Compared with LQR control, triple inverted pendulum based on ANFIS control has better dynamics performance and anti-interference capability.