The global greenhouse gas emitted from shipping activities is one of the factors contributing to global warming; thus, there is an urgent need to mitigate the adverse effect of climate change. One of the key strategies is to build a vibrant maritime industry with the use of innovation and digital technologies as well as intelligent systems. The digitization of the shipping industry not only provides a competitive edge to the shipping business model but also enhances ship operational and energy efficiency. This review paper focuses on the big data analytics and machine learning applied to harbour craft vessels with the aim to achieve fuel efficiency. The paper reviews the telemetry system requires for the digitalization of harbour craft vessels, its challenges in installation, the vessel monitoring and data transmission system. The commonly used methods for data cleaning are also presented. Last but not least, the paper considers two types of the machine learning systems, i.e., supervised and unsupervised machine learning systems. The multi-linear regression and hidden Markov model for supervised machine learning system and the artificial neural network, grey box model and long short-term memory model for unsupervised machine learning are discussed, and their pros and cons are presented.
This paper presents the analysis to reduce carbon emission from tugboat operations by utilizing a proposed unsupervised machine learning operational scoring system. The time-series analysis is performed by transforming data into a common domain for clustering. The data are collected from a tugboat to investigate the correlation between environmental and location data with fuel consumption to achieve fuel efficiency. The relevant parameters that influence the fuel consumption of the tugboat, such as fuel consumption, vessel route, vessel speed and wind metrics are collected from sensors installed onboard the ship and data provider to monitor and to gauge the vessel’s performance. The raw readings are conditioned (data cleaning and data pre-processing) before transformation to Score Dataset: the Raw mass-flowrate readings are cleaned by using the Haar wavelet; the wind raw reading is converted to wind effect data; the Location data is converted to vessel speed data. Together, they form a Score Dataset by applying the time series K-means clustering. The subsequent unsupervised learning identifies the activity labels that describe qualitatively the operations of the vessels and are obtained by using the non-time series K-mean clustering. By using the Hidden Markov Model approach, this paper attempts to explain the stochastic correlation among parameters explained earlier. The correlation is the information of newly discovered knowledge in terms of likelihood matrices, also known as the knowledge base (KB). The KB may be consumed to perform predictions. Hence, it is possible to suggest the optimal ship operation, i.e., speed that produces the optimum fuel consumption. The Score Dataset and clustering that are produced in this paper could also be used in the Artificial Neural Network for future work.
Harbor craft historical routes contain valuable information on how the experienced crews navigate around the known waters while performing jobs. The noon report logs each job timeframe which can be used to segregate the time-series positional data as routes. Other information from the noon report such as fuel consumption could be associated with a particular job as well. This paper offers a solution to encompass crew navigational experience into neural network models. The variational autoencoder, which is a generative model, can capture the routes into a knowledge base model. The same variational autoencoder is also able to train other neural networks to make predictions of route and fuel consumption based on job metadata (I.e., job duration, activity area, and route classification). The predicted routes could be used as a cost map for pathfinding algorithms such as A* or Dijkstra.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.