摘要
为提高微小型水下航行器运动控制的机动性和避障能力,提出一种广义S型模糊神经网络(SFNN)控制方法.采用广义Sigmoid函数作为隶属函数,并推导出基于最小扰动的网络学习方法补偿敏感性.与Gauss型模糊神经网络(FNN)进行比较并以"微龙"号水下航行器为研究对象进行了试验研究.结果表明,采用广义SFNN控制,在没有损失整体控制品质和稳定性的情况下,控制系统响应速度大幅度提高,反应能力增强,从而满足微小型水下航行器的实时控制要求.
A novel control method based on generalized Sigmoid fuzzy neural network （SFNN） was developed to improve maneuverability and ability of avoiding obstacles in motion control of mini underwater vehicles. The structure of SFNN was constructed according to motion characters, and generalized Sigmoid function was selected as membership function. The learning algorithm which calculated dynamic learning ratio based on least disturbance was deduced to compensate sensibility. Comparison with Gauss fuzzy neural network（FNN） was done and experiments were carried out on ＂Weilong＂ mini underwater vehicle. The results show that the respond speed of control system increases greatly without loss of integral control quality and stability, and response ability is improved so as to meet real-time control requirement of mini underwater vehicles.
出处
《大连海事大学学报》
EI
CAS
CSCD
北大核心
2007年第3期11-15,共5页
Journal of Dalian Maritime University
基金
国家自然科学基金资助项目(50579007)
