The real-time provision of high-quality estimates of the ocean wave parameters at appropriate spatial resolutions are essential for the sustainable operations of marine structures. Machine learning affords considerable opportunity for providing additional value from sensor networks, fusing metocean data collected by various platforms. Exploiting the ship-as-a-wave-buoy concept, this article proposes the integration of vessel-based observations into a wave-nowcasting framework. Surrogate models are trained using a high-fidelity physics-based nearshore wave model to learn the spatial correlations between grid points within a computational domain. The performance of these different models are evaluated in a case study to assess how well wave parameters estimated through the spectral analysis of ship motions can perform as inputs to the surrogate system, to replace or complement traditional wave buoy measurements. The benchmark study identifies the advantages and limitations inherent in the methodology incorporating ship-based wave estimates to improve the reliability and availability of regional sea state information.
Digital Twins have much attention in the shipping industry, attempting to support all phases of a vessel’s life cycle. With several tools appearing in Digital Twin software suites, high-quality manoeuvring and performance prediction remain cornerstones. Propulsion efficiency is in focus while in service. Simulator-based training is in focus to ensure safety of manoeuvring in confined waters and harbours. Prediction of ships’ velocity and turn rate are essential for correct look and feel during training, but phenomena like dynamic inflow to propellers, bank and shallow water effects limit simulators’ accuracy, and master mariners often comment that simulations could be in better agreement with actual behaviours of their vessel. This paper focuses on digital twin enhancements to better match reality. Using data logged during in-service operation, we first consider a system identification perspective, employing a first-principles model structure. Showing that a complete firstprinciples model is not identifiable under the excitation met in service, we employ a Recurrent Neural Network to predict deviations between measured velocities and the model output. The outcome is a hybrid of a first-principles model with a machine learning generic approximator add-on. The paper demonstrates significant improvements in prediction accuracy of both in-harbour manoeuvring and shallow water passage conditions.
This paper presents an assessment of three methods used for sea state estimation via the wave buoy analogy, where measured ship responses are processed. The three methods all rely on Machine Learning exclusively but they have different output; Method 1 provides bulk parameters, Method 2 yields a point wave spectrum and the wave direction, while Method 3 gives the directional wave spectrum in non-parametric form. The assessment is made using full-scale data from an in-service container ship in cross-Atlantic service. Training and testing of the methods are made using data from a wave radar, and the three methods perform well. An uncertainty measure, equivalently, a trust level indicator, based on the variation between the post-processed outputs of the methods is proposed, and this facilitates determination of estimates with small errors; without knowing the ground truth.
Abstract: In recent years, the development of ground robots with human-like perception capabilities has led to the use of multiple sensors, including cameras, lidars, and radars, along with deep learning techniques for detecting and recognizing objects and estimating distances. This paper proposes a computer vision-based navigation system that integrates object detection, segmentation, and monocular depth estimation using deep neural networks to identify predefined target objects and navigate towards them with a single monocular camera as a sensor. Our experiments include different sensitivity analyses to evaluate the impact of monocular cues on distance estimation. We show that this system can provide a ground robot with the perception capabilities needed for autonomous navigation in unknown indoor environments without the need for prior mapping or external positioning systems. This technique provides an efficient and cost-effective means of navigation, overcoming the limitations of other navigation techniques such as GPS-based and SLAM-based navigation. Graphical Abstract: [Figure not available: see fulltext.]
Harbour vessel emissions are growing concerns in the maritime industry regarding environmental sustainability. Accurate emissions prediction can stand in monitoring and addressing the issue. This study proposes a machine-learning approach using Artificial Neural Network (ANN) for predicting harbour vessel emissions. The approach shows superiority over the bottom-up method introduced by the 4th IMO GHG Study regarding prediction accuracy. Actual emissions data from onboard measurements are used for training ANN models and as references for evaluating the methods. Compared to the bottom-up method, the improvement in error reduction can be up to 30% for predicting nitrogen oxides and 54% for carbon monoxide when only using ship-related factors as input variables. By adding selected meteorological factors in the experiments, the prediction accuracy enhancement can achieve up to 48% for nitrogen oxides and 62% for carbon monoxide. The proposed ANN approach could assist relevant stakeholders in improving emissions prediction and operations optimisation.