Finding safe and efficient shipping routes in ice-covered waters: A framework and a model

• Published in: Cold regions science and technology Vol. 165; p. 102795
• Main Authors: Lehtola, Ville, Montewka, Jakub, Goerlandt, Floris, Guinness, Robert, Lensu, Mikko
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.09.2019
• Subjects: Geo-information; Ice navigation maritime safety; Multi-objective optimization; Pathfinding; Route optimization; Pathfinding; Multi-objective optimization; Ice navigation maritime safety; Geo-information; Route optimization
• ISSN: 0165-232X; 1872-7441; 1872-7441
• DOI: 10.1016/j.coldregions.2019.102795

Safety for conventional and autonomous navigation in ice-covered waters is a topic of rising importance. Here, we propose a generic extendable framework to provide the optimal route from multiple route planning objectives. These objectives are attained by an evaluation of multi-source input data, including state-of-the-art model data for ice conditions, for bathymetric knowledge, and for ship-ice interaction. Additionally, we model the ship-ship interactions statistically using a mean-field, to account for ships (indirectly) assisting each other via artificial ice channels. For the subsequent pathfinding problem, we propose a new A*-based algorithm that yields output which is not dependent on the grid format of the input data but instead consists of a path that follows the Earth's curvature. The outputs of the algorithm are a set of waypoints (representing the optimal route), the travel costs (expressed in time), and the additional travel cost estimates caused by route deviation, should the optimal route be altered in any way. The steaming speeds, the optimal route, and the deviation times are represented with two-dimensional (2D) maps. Finally, we provide a model implementation of our framework as a Matlab-package, ICEPATHFINDER, that is suitable for both operational and strategic ship route optimization. •Our framework can incorporate multiple data sources and pathfinding objectives for route planning in ice covered waters.•Two novel map representations are introduced for a given ship class in given ice conditions: a speed map, and a time map.•The indirect ship-ship interaction via artificial ice channels is incorporated via a mean-field model using historic AIS data.•Our Matlab implementation, ICEPATHFINDER, is used to output physically realistic geodesic paths on the Baltic Sea.