Dynamic Positioning Control of Unmanned Surface Vehicle for Core Sampling; Correction Input Using Neural Network

Fujii, Shun*; Tahara, Junichiro*; Baba, Shoichiro*; Sanada, Yukihisa   

We have been studying an USV for measuring radiation in the sea near the Fukushima Daiichi Nuclear Power Plant. The USV is equipped with a core sampler and a sensor to measure radiation on the seafloor, and it is necessary to maintain the position of the USV hen using these devices. In the sea area where the USV is used, there are disturbances such as wind, current, waves, which make it difficult to maintain the position. To solve this problem, we devised a position control method that dynamically adds correction inputs using a neural network.