融合卡尔曼滤波的无人船航向神经网络控制

Heading neural network control of unmanned surface vessel with Kalman filter

为实现无人船的航向自主跟踪功能,考虑随机性环境扰动对船舶运动的影响,提出一种融合卡尔曼滤波器的自适应神经网络航向分数阶控制方法。对无人船的航向跟踪控制系统和外部环境干扰进行数学建模,设计该模型下的扩展卡尔曼滤波器,避免环境因素对船舶操纵性能的影响,基于分数阶微积分理论设计无人船航向控制的分数阶控制器,给出一种径向基函数神经网络学习算法,用于解决该控制器的参数整定问题。对所提方法进行仿真实验,实验结果表明,该方法能使无人船快速稳定地跟踪期望参考航向,具有良好的自适应性和鲁棒性。

To realize the autonomous heading tracking function of unmanned surface vessel,the influence of random environmental disturbance on ship motion was taken into consideration,and an adaptive neural network heading fractional order control method with Kalman filter was proposed.The mathematical models of unmanned surface vessel heading tracking control system and external environment disturbance were established,an extended Kalman filter was designed to avoid the influence of environmental factors on ship steering performance.A heading fractional order controller based on fractional calculus theory of unmanned surface vessel was designed,and a learning algorithm of radial basis function neural network was proposed to tune controller parameters.The simulation experiment of the proposed method was carried out.The results show that the proposed method can make the unmanned surface vessel track the desired reference heading quickly and steadily,which has good adaptability and robustness.