基于神经网络模糊控制理论的转台伺服系统控制设计

Control of Turntable Servo System Based on NNSOC Method

针对三轴飞行仿真转台伺服系统非线性、模型不精确等特点,在分析转台系统结构的基础上,采用模糊PID控制的方法对转台伺服系统进行仿真控制,得到较好的控制效果;模糊控制控制规则的获得带有很大的人为因素,并且在控制过程中对规则采用查表法占用大量的内存;基于以上原因,设计了神经网络模糊控制器(NNSOC),利用神经网络控制自学习、自调整的能力,为模糊控制器提供自动生成控制规则的能力;同时由于神经网络具有联系记忆能力,可对未训练的样本做出决策;对NNSOC的控制效果进行了仿真;结果表明:其具有很好的动态性能和鲁棒性,对转台的控制效果良好。

The three-axis turntable is nonlinear and the model is not precise.The paper analyses the structure of the turntable and adapts Fuzzy PID control method to control the turntable servo system.Also,make some simulation and result shows that the effect of Fuzzy PID is good.However,the rules of Fuzzy PID controller get from the experience of technician and take too much memory space.For these reasons,the paper designs a Neural Network （NN） Self-organizing Controller （NNSOC）.Using the self-study and self-adjust ability of Neural Network,NNSOC provides control rules to Fuzzy PID controller.Meanwhile,because of the memory ability of NN,NNSOC can make decisions to the untrained samples.Make some simulation on NNSOC.Results show that NNSOC has good abilities of and robustness,the effect of controlling on turntable servo system is good.