A STUDY ON SHIP MANEUVERABILITY ESTIVATION AND AUTOMATIC NAVIGATION BASED ON A PREDICTIVE CONTROL

本课题的研究目的是开发船舶位置控制和自动导航系统。但要开发一个综合考虑多方面因素的船舶位置控制系统仍然是一个难点。因此，本文从港口航行特别是靠泊这一对航海员来说是一种负担的情况出发，在这种情况下，自动定位的导航支持系统是非常有效的，通常通过系留模型试验来评估船舶操纵性。但目前的模型试验技术中，低速试验比较困难。所以我们做了一个自由运行的测试，而不是一个captivemodel测试。首先进行了真实的实船试验。自激试验在分析中存在的问题被指出后，一直没有得到重视。在真实的实船试验中，我们利用高精度的卫星定位系统--动态GPS，开发了试验系统。在船模试验中，我们研制了一套由视频定位分析仪和风机组成的船模运动实验系统，模拟拖船的运动。在船模位置控制方面，我们研究了神经网络控制系统，并对靠泊进行了仿真，设计了自动靠泊控制系统。最后，探讨了将此位置控制系统装备成导航辅助系统的可能性。

The aim of This study is development of ship position control and automatic navigation system. But it is still difficult to develop a ship position control system considered many aspects. So we started this study from the case of navigation in harbor especially berthing, that is a burden to a navigator, and in that case the navigation supporting system by automatic positioning is vary effective.Captive model test was usually carried out to estimate a ship maneuverability. But a low speed test is difficult in present captive model test technique. So we made a free-running test instead of a captivemodel test. Firstly we carried out a real ship test. Free-running test was not attached importance, since its problem in analysis was pointed out. We tried to break through by devising test items and developing a statistical analysis utilizing database.In a real ship test, we developed experiment system by using a kinematic GPS, which is a satellitic positioning system with high accuracy. In a model test, we developed a experiment system with a video positioning analyser and air fans for moving model ship supposed tug boats.About ship position control, we investigated neural network control system and made a simulation for a berthing, and designed automatic berthing control system. Next subject is generalization to many aspects.Finally, we examined possibility to equip this position control system into a navigation supporting system.