CAJ | 학술논문

针对船舶航向控制中模型参数变化引起的不确定性问题,提出一种基于动态神经模糊模型(DNFM)的自适应控制算法.DNFM在学习中同时调整结构和参数,充分逼近Norrbin非线性船舶NARX模型的逆动力学.训练好的DNFM与PD控制器并联构造自适应控制器用于航向控制,且以PD控制器的输出作为自适应律的自变量,进一步在线调整DNFM的权值.以5446 TEU集装箱船的航向控制仿真结果验证了算法的有效性.
침대선박항향공제중모형삼수변화인기적불학정성문제,제출일충기우동태신경모호모형(DNFM)적자괄응공제산법.DNFM재학습중동시조정결구화삼수,충분핍근Norrbin비선성선박NARX모형적역동역학.훈련호적DNFM여PD공제기병련구조자괄응공제기용우항향공제,차이PD공제기적수출작위자괄응률적자변량,진일보재선조정DNFM적권치.이5446 TEU집장상선적항향공제방진결과험증료산법적유효성.
A dynamic neural fuzzy model (DNFM) based adaptive control algorithm for ship course control was developed to overcome uncertainties arising from changes of model parameters. The DNFM identified the inverse dynamics of the Norrbin nonlinear auto regressive exogenous-inputs (NARX) ship model sufficiently, while the structure and parameters themselves were adjusted simultaneously. Well-trained DNFM was then connected with a proportional-plus-derivative (PD) controller in parallel to construct an adaptive controller for course control. The weights of the DNFM were further adjusted by an adaptive law whose independent variable was the output of the PD controller. Simulation results for course control of a standard 5446 twenty-foot equivalent unit (TEU) container vessel validated the effectiveness of the proposed algorithm.