基于CMAC的中性浮力实验体解耦控制研究  

Decoupling Control Research of Neutral Buoyancy Plant Based on CMAC

作　　者：李佩龙[1] 王磊[1] 李高丰[1] 

机构地区：[1]同济大学电子与信息工程学院,上海201804

出　　处：《系统仿真学报》2015年第5期1003-1009,共7页Journal of System Simulation

基　　金：国家863高科技计划(2012AA0392)

摘　　要：针对中性浮力实验体控制问题,提出一种动态解耦多通道复合控制算法。通过实验体受力分析,建立了其六自由度运动模型。设计了一种动态解耦神经网络,其优点在于能适应动态变化的耦合程度。采用CMAC(Cerebellar Model Articulation Controller)和PID复合控制进行单通道控制,其优点在于结合了CMAC前馈控制的快速响应性能和PID反馈控制的强抗干扰性。经MATLAB仿真验证,提出的算法能有效控制实验体位置和姿态。调整时间短且稳态精度高,效果优于PID控制。To solve the control problem of neutral buoyancy plant, a dynamic decoupling and multidimensional compound control agorithm was proposed. By the plant force situation analysis, its six degree motion control model was developed. Then a dynamic decoupling neural network was designed with the merit of adaptation of the dynamic changing coupling situation. CMAC(Cerebellar Model Articulation Controller) and PID compound control was adopted in single channel control, whose advantage is that it combines the rapidity of CMAC feedforward control and anti-interference of PID feedback control. MATLAB simulation results show that the algorithm can control the plant's position and posture effectively. The agorithm has short adjusting time and high stability precision, which outpeforms PID control.

关 键 词：中性浮力 CMAC PID 神经网络解耦 

分 类 号：V416.5[航空宇航科学与技术—航空宇航推进理论与工程] TP273[自动化与计算机技术—检测技术与自动化装置]