摘要
针对常规水面船舶的航向跟踪控制中建模参数随航速时变引起的不确定性问题,提出一种基于动态神经模糊模型的控制算法.动态神经模糊模型在学习中同时调整结构和参数,充分逼近船舶的逆动力学.训练好的动态神经模糊模型作为逆控制器,与传统的PD控制器并联,用于船舶航向的跟踪控制,且控制过程中能进一步调整模型权值.以5446TEU大型集装箱船为例,进行航向跟踪控制的仿真结果表明,该算法能克服建模参数不确定性的影响,快速有效地跟踪期望航向,控制效果良好.
Aiming at the uncertainties arising from modeling parameters changing with the ship speed in course tracking control,a dynamic neuro-fuzzified model（DNFM） based tracking control algorithm is presented in this paper. The DNFM is sufficiently identified with the inverse dynamics of ship, while the structure and parameters are adjusted simultaneously. Then the well-trained DNFM, as an inverse controller,is parallel-connected with a conventional PD controller for course tracking control of the ship, with the weights value of DNFM further adjusted in this process. Simulation results of the course tracking of the large-type 5446TEU container show that the algorithm can track the course desired quickly with good control results by overcoming the effect from certainties of the modeling parameters.
出处
《南京信息工程大学学报（自然科学版）》
CAS
2009年第3期218-222,共5页
Journal of Nanjing University of Information Science & Technology（Natural Science Edition）
基金
国家自然科学基金(60774046)
关键词
航向跟踪
不确定性
动态神经模糊模型
规则调整
course tracking
uncertainties
dynamic neuro-fuzzified model
rule adjustment
