Modeling and Analysis of Motion Data from Dynamically Positioned Vessels for Sea State Estimation

Abstract

Developing a reliable model to identify the sea state is significant for the autonomous ship. This paper introduces a novel deep neural network model (SeaStateNet) to estimate the sea state based on the ship motion data from dynamically positioned vessels. The SeaStateNet mainly consists of three components: an Long-Short-Term Memory (LSTM) recurrent neural network to capture the long dependency in the ship motion data; a convolutional neural network (CNN) to extract time-invariant features; and a Fast Fourier Transform (FFT) block to extract frequency features. A feature fusion layer is designed to learn the degree affected by each component. The proposed model is applied directly to the raw time series data, without needing of any hand-engineered features. A sensitivity analysis (SA) method is applied to assess the influence of data preprocessing. Through benchmark test and experiment on ship motion dataset, SeaStateNet is verified effective for sea state estimation. The investigation on real-time test further shows the practicality of the proposed model.

A Novel Densely Connected Convolutional Neural Network for Sea-State Estimation Using Ship Motion Data

“…Although this method does not rely on accurate mathematical models, it requires a lot of human involvement. To reduce the influence of artificial features, Cheng et al proposed a deep learning based end-to-end model for sea state estimation using the DP motion data [3]. While datadriven approaches have had good results, these approaches have not considered wave direction.…”

“…[30]. Most researches generally define the world-wide sea state by wave height, as shown in TABLE I. Data-driven approaches often label the sea state based on TABLE I [3], [7]. However, this labeling approach ignores the information of wave direction.…”

“…• SeaStateNet: SeaStateNet is specifically designed for sea state estimation [3]. SeaStateNet is composed of three parallel parts: LSTM part, CNN part, and FFT part.…”

“…Traditional technologies to estimate the sea state usually include manual observations, wave buoys, X-band wave radar, and meteorological remote sensing satellites or weather forecasts [3]. The advantage of manual observations is that the data has high persistence and does not depend on external sensors.…”

“…Nevertheless, to the best of our knowledge, the previous model-free methods only considered the onboard measurement of dynamic positioning (DP) motion and only considered the height of the waves without considering the direction of the waves [3], [7]. DP motion, used in [3], [7], represents a special kind of maneuvering, which involves maintaining a fixed location or performing a very slow tracking task [8]. The use of this special maneuvering to estimate sea state lacks generality because most ships do not have a DP system, and those that do are generally moving forward when in operation.…”