摘要
三体滑行艇的航态问题是一个极其重要的研究内容,其具有预测难度大、受力复杂、稳定性差等特点。目前,DT决策树以及LSTM(Long Short-Term Memory networks)被广泛应用于故障排除、无人控制、路径规划等方面,并具有较好的工程适用性。在本文中,为了研究三体滑行艇在各航速状态下的航态问题,利用DT决策树+LSTM算法,根据三体滑行艇的运动姿态数据进行训练学习,得到艇体各运动参数的权重分配、实时预测,构建艇体运动姿态数据库,实现对于三体滑行艇的航速调节与匹配。通过试验数据与数值仿真数据的对比分析,得出该方法对于三体滑行艇航态的预测具有较高的准确性,设计所得航速控制器对三体滑行艇在各航态下的航速匹配具有较好的工程适用性。
The navigational problem of the trimaran planing craft is always an extremely important research content,which has the characteristics of large prediction difficulty,complex forced situation and poor stability. Currently,DT(Decision tree)and LSTM(Long Short-Term Memory networks)algorithm are widely used in speech recognition,image processing,path planning,machine translation,etc.,and have good engineering applicability. In this paper,in order to study the navigation state of the planing trimaran in each speed state,the DT+LSTM algorithm is used to carry out the training and learning according to the motion attitude data of planing trimaran,and the weight distribution of each motion parameter of the hull is obtained. Real-time prediction is performed and a hull movement attitude database is achieved to adjust the craft speed to match to its motion attitude.Through the analysis of experimental data and numerical simulation data,it is concluded that the method has high accuracy for the prediction of the planing trimaran’navigational state. The designed speed controller has good engineering applicability to the speed matching of the trimaran planing craft in each navigation state.
作者
侯永康
邹劲
董超
刘蔚
谈果戈
HOU Yongkang;ZOU Jin;DONG Chao;LIU Wei;TAN Guoge(College of Shipbuilding Engineering,Harbin Engineering University,Harbin 150001;South China Sea Survey Technology Center,State Oceanic Administration,Key Laboratory of Marine Environmental Survey Technology and Application,Ministry of Natural Resources,Guangzhou 510000)
出处
《舰船电子工程》
2020年第1期40-45,98,共7页
Ship Electronic Engineering
基金
重点实验室基金项目（编号：614222303030917）
全球变化与海气相互作用专项资助
