Abstract

We are developing a small ASV (μ-ASV) with a length of 2 m based on a surfboard as one of the small marine research instruments. The μ-ASV will be used as a mother ship of the small ROV. In this case; it is difficult to model and design the control of the μ-ASV in advance because the motion characteristics of the μ-ASV change from time to time due to the tension of the ROV cable. Therefore; in this study; the modeling of the μ-ASV is performed in real time using LSTM; a kind of neural network.

Introduction

Recently; there has been a lot of research on oceanographic surveys using small oceanographic instruments (Bai; 2020; Casas; 2019; Conte; 2020; Karapetyan; 2018). We are developing a small ASV (μ-ASV; Fig. 1) as one of such small oceanographic survey instruments (Baba; 2019; Baba; 2020). The μ-ASV is based on a surfboard and has a length of 2 m and an airborne weight of 50 kg; which is small enough to be carried by a minivan and can be operated by two or three people. Currently; the μ-ASV can measure the depth and sediment quality and take topographic images of the seafloor using side-scan sonar. In this case; the μ-ASV can set a course and follow the course.

In addition to the above; we plan to use the μ-ASV as a mother ship for a small ROV (Fig. 2). In this case; the μ-ASV is expected to automatically move to a target point with the ROV on board; throw the ROV into the sea; and control the ROV and cables while maintaining a fixed point. However; the motion characteristics of the μ-ASV change from time to time when the above operation is performed. The reason for this is the effect of the tension of the cable generated by the movement of the ROV after it is put into the sea; and the effect of the tide and wind.

This content is only available via PDF.
You can access this article if you purchase or spend a download.