基于天牛须BP神经网络的船舶电动舵机自抗扰控制

THE ACTIVE-DISTURBANCE-REJECTION CONTROL OF MARINE ELECTRIC STEERING GEAR BASED ON LONGHORN WHISKER BP NEURAL NETWORK

将自抗扰控制器(ADRC)引入到智能船舵机控制优化问题中,同时将神经网络嵌入到自抗扰控制,针对电动舵机快响应、非线性的特点提出一种神经网络自抗扰控制。将天牛须搜索(BAS)寻优能力强的特性与BP神经网络中的权值与阈值寻优过程相结合,建天牛须BP神经网络自抗扰控制器,弥补了BP神经网络训练时间较长、收敛速度较慢的问题。以船舶电动舵机系统作为被控对象在该控制器下的舵机系统的控制性能得到了提升,控制精度高无超调同时具有良好的鲁棒性,同时建立的BAS-BP神经网络自抗扰控制器对于实际应用具有指导意义。

The active disturbance rejection controller(ADRC)was introduced into the intelligent ship steering gear control optimization problem,and the neural network was embedded in the ADRC.Aimed at the characteristics of fast response and nonlinear electric steering gear,a neural network active-disturbance rejection control was proposed.Combining the characteristics of BAS with the weight and threshold optimization process of BP neural network,an active disturbance rejection controller of BAS neural network was built to make up for the long training time and slow convergence speed of BP neural network.Taking Marine electric steering system as the controlled object,the control performance of the steering system under this controller was improved,the control precision was high,with no overshoot and has good robustness.The established BAS-BP neural network active disturbance rejection controller had guiding significance for practical application.