虚拟锚泊神经网络模糊滑模控制器的设计

Design of A Neural Network Fuzzy Sliding Mode Controller for Virtual Mooring

针对虚拟锚泊浮标在目标点位置难以进行锚泊的问题,设计了一种神经网络模糊滑模控制器。通过引入虚拟锚泊圆,将在锚泊点一定范围内虚拟锚泊的问题转变为对虚拟锚泊圆进行路径跟踪的问题,应用Fossen矩阵建立虚拟锚泊浮标动力学模型,采用Serret-Frenet坐标变换和重新定义输出变量,将单输入多输出非线性系统转变为单输入单输出非线性系统,进而构建滑模控制器,采用模糊算法对滑模开关量进行在线调整,削弱滑模控制器的'抖振',采用神经网络对滑模控制器进行优化,摆脱对系统模型的依赖性,提高控制器的鲁棒性,并对该控制器进行仿真实验和分析,结果表明该控制器具有很好的虚拟锚泊性能。

Aiming at the problem that the virtual mooring buoy was difficult to moor at the target location,a neural network fuzzy sliding mode controller was designed in this paper.The virtual mooring circle was introduced to convert the problem on virtual mooring within a certain range of the mooring point into the problem on path tracking for the virtual mooring circle.The Fossen matrix was used to build the dynamic model of the virtual mooring buoy.The Serret-Frenet coordinate transformation and redefinition of output variables were utilized to convert single-input multi-output nonlinear system into single-input single-output nonlinear system and a sliding mode controller was constructed.The chattering phenomenon of the siding mode control was alleviated and adjusted by the fuzzy algorithm in real time.The neural network was used to optimize the sliding mode controller,get rid of the dependency of the system model and improve the robustness of the controller.Simulation and analysis on the controller were carried out,the results showed that the proposed controller had a good performance of virtual mooring.