基于模糊神经网络的舵鳍联合减摇控制

Rudder-Fin Joint Roll Damping Control Based on Fuzzy-Neural Network

船舶是典型的复杂非线性、不确定性系统,受自身结构、所装货物质量、航速及风浪流等因素的影响,难以建立其精确的数学模型。由此,针对目前舵鳍联合减摇控制设计依赖于船舶模型参数的缺点,在充分考虑舵角、舵速等约束条件的基础上,基于模糊神经网络设计分离型舵鳍联合减摇控制器,并在不同海况和模型参数存在摄动的情况下进行MATLAB仿真研究。仿真结果表明:所设计的模糊神经网络控制器不仅能使船舶保持航向,而且其减摇效果优于传统PID控制,同时体现出了极强的鲁棒性。

Ships are typical complex nonlinear and uncertain systems influenced by ship structure itself, loading, speed and the wave interference, therefore, difficult to accurately model. Present ship parameter dependent design of rudder/fin joint damping system is not able to cope with such problem, so a rudder-fin joint roll damping controller separated from the ship is designed based on the fuzzy-neural network principle and the full consideration of the constraints of the rudder angle and steering speed. Simulations are carried out under different sea condition and various perturbations by MATLAB. The simulation results prove that compared with PID controller, the fuzzy-neural network controller performes better in maintain ship course and damping rolling with strong robustness.