Sugeno fuzzy control adaptive fuzzy neuralnetwork gradient d

  本文关键词：基于Takagi-Sugeno模糊神经网络的欠驱动无人艇直线航迹跟踪控制，由笔耕文化传播整理发布。

基于Takagi-Sugeno模糊神经网络的欠驱动无人艇直线航迹跟踪控制

Straight-path tracking control of underactuated USV based on Takagi-Sugeno fuzzy neural network

[1] [2] [3] [4] [5] [6]

Dong Zaopeng, Liu Tao , Wan Lei, Li Yueming, Liao Yulei, Liang Xingwei（1. National Key Laboratory of Science and Technology on Autonomous Underwater Vehicle, Harbin Engineerin

[1]哈尔滨工程大学水下机器人技术重点实验室,哈尔滨150001; [2]哈尔滨工程大学船舶工程学院,哈尔滨150001

文章摘要：研究一类欠驱动无人艇的直线航迹跟踪控制问题,提出了一种自适应T-S（Takagi-Sugeno）模糊神经网络控制方法。首先在神经网络体系结构中设计前件网络匹配T-S模糊控制器的模糊规则前件,设计后件网络进行T-S模糊运算推理从而生成模糊规则后件;其次基于梯度下降法原理,设计了T-S模糊规则参数的优化学习算法;然后结合BP神经网络的误差反向传播原理和梯度下降法,设计了模糊神经网络体系误差的反向传播迭代算法,用于高斯隶属度函数参数的学习优化;最后设计了基于T-S模型的模糊神经网络控制器,并通过仿真实验验证了所提出方法和所设计控制器的有效性。

Abstr：The straight-path tracking control for a class of underactuated unmanned surface vehicle （USV） is discussed and an adaptive T-S （Takagi-Sugeno） fuzzy neural network control method is proposed in this paper. Firstly, the antecedent network of the neural net- work architecture is designed to match the antecedent of fuzzy rules for T-S fuzzy controller; while the consequent network is designed to generate the consequent of fuzzy rules with T-S fuzzy inference. Secondly, the optimization learning algorithm for the parameters of T-S fuzzy rules is designed based on the gradient descent method. Thirdly, the back propagation iterative algorithm for the error of fuzzy neu- ral network system is designed based on the back propagation principle of BP neural network and gradient descent method, which is used for the learning and optimization of the Gaussian membership function parameters. At last, the fuzzy neural network controller based on T-S model is designed. Simulation experiment was conducted, and the results verify the effectiveness of the proposed control method and the designed controller.

文章关键词：

Keyword：:underactuated unmanned surface vehicle （USV） straight-path tracking Takagi-Sugeno fuzzy control adaptive fuzzy neuralnetwork gradient descent method back propagation iterative algorithm

课题项目：国家高技术研究发展计划（863计划）（2012AA09A304;2014AA09A509）; 国家自然科学基金（51409054;51409059;51409061）; 中国博士后科学基金（2013M540271）项目资助

  本文关键词：基于Takagi-Sugeno模糊神经网络的欠驱动无人艇直线航迹跟踪控制，，由笔耕文化传播整理发布。