Innovative Approaches to Support Precision Navigation

Nautical charts are compiled from geospatial information of varying quality, collected at different times, using various techniques. In maritime navigation failure to take chart data quality into account can be one of the factors leading to maritime accidents. The first approach of the hydrographic community for informing mariners about the data quality on charts was with a description in the title of the chart, which with time took the form of a chart inset either with the use of the source diagram or with the more complex reliability diagram. In the early 1990s, the hydrographic community introduced the Category of Zones of Confidence (CATZOC) for use on paper and the newly introduced Electronic Navigational Charts (ENCs). The Quality of Bathymetric Data (QoBD) is the newest development for use in the S-100 series of standards. In QoBD, the five ZOC alphanumeric categories of assessed data A1, A2, B, C, and D are renamed to 1, 2, 3, 4, and 5, respectively. One more category “O” (Oceanic) is provided for the areas where water depth is deeper than 200m and, thus, does not pose a threat to surface navigation. The horizontal and vertical uncertainties and the seabed coverage criteria for each category remain unchanged, while an attribute for the temporal variation of the seabed is added. Despite these changes, however, the legibility and utility of the current methods are limited, and therefore the aim of this research project is the development of new visualization and integration methods of bathymetric data quality in ECDIS in support of decision making on board.

In the 2019 reporting period, Christos Kastrisios, Colin Ware, Brian Calder, and Tom Butkiewicz, in collaboration with Lee Alexander and Rogier Broekman, reviewed the deficiencies of the current CATZOC symbology and integration in route planning and execution. Subsequently, the research team studied recent research into the portrayal of bathymetric data uncertainty and set the requirements that the new visualization method should satisfy to be effective for the application.

Accordingly, in the 2020 reporting period, Christos Kastrisios and Colin Ware considered how different visual variables might be used to meet the requirements and proposed the use of a sequence of textures created by combining two or more visual variables. Two countable textures schemes were developed: one consisting of lines (Lines) and one consisting of clusters of dots (Dot-Clusters) with the fundamental principle being that the number of lines or dots represent the QoBD. These were evaluated in 2021 reporting period with a user online survey and in-lab experiment along with three others, two of them adopting ideas previously expressed in the hydrographic community (one with opaque color fills (Opaque-Colors) and one of transparent color fills (Transparent-Color)) as well as see-through color textures (Color-Textures) scheme in an effort to account for the color blending issues of the former two.

Figure 32-1. An example of the lines scheme using the developed web-service.

The two proposed schemes of countable textures demonstrated the best performance in the survey and experiment combined, followed by the opaque-colors. In the 2022 Reporting period, the research team published the results of the on-line survey and in-lab experiment and presented them at IHO relevant working groups (WG) meetings (i.e., Data Quality WG, Nautical Cartography WG, and S-101 Project Team meetings). The S-101 project team tasked the research team to translate the two textures into an appropriate format for use with ECDIS as well as fine tuning them so that they become part of the S-100 IHO Portrayal registry.

To this end, in the 2023 reporting period, with Paul Johnson, and Craig Greene and Tom De Puyt from ESRI, the team set up an online Maritime service, hosted on CCOM server, for testing the new symbology. An example of our lines scheme using the web-service is illustrated in Figure 32-1. Besides this effort, the web-service aims to facilitate the development, testing, and sharing of results with the hydrographic community of other CCOM efforts such as our on-going Cartosemiotics effort (Task 37).

To ensure that the developed SVGs are S-100 compliant, with the support of David Grant from US Navy, we migrated to the S-100 viewer for testing the developed SVGs. Figure 32-2 is an example of an S-100 Test Dataset with Quality 3 with Lines in Day and Dot Clusters in Dusk mode.

Figure 32-2. S-100 Test Dataset with Quality 3 with Lines in Day (left) and Dot Clusters in Dusk mode (right) in the S-100 viewer.

The Lines and Dot-clusters SVGs in all three ECDIS modes (Day, Dusk, and Night) were delivered to S-101 Project Team and became part of the IHO Portrayal Catalogue Alternate in late September. IHO S-101 PT expects IHO-Singapore lab to test the symbology with S-100 compliant ECDIS in simulators and at sea along with other new symbology incorporated in the newest version of the alternate Portrayal Catalogue.

The above work on new symbology to support the representation of QoBD is expected to make it easier for mariners to understand the uncertainty of bathymetric information, however, reasoning about horizontal and vertical uncertainty can still be challenging. We have therefore been investigating new ways of representing and integrating bathymetric uncertainty in Electronic Charts. In the 2023 reporting period, the research team revisited the aspect of data quality integration in ECDIS, last reported in 2019, and particularly the decision tree for the “safe route” approach.

New Methods to Support Route Planning with Uncertainty
Even when uncertainty information is clearly displayed, the process of reasoning about which potential passages are safe can impose a significant cognitive burden. The route planner must consider which points and contours might be potential hazards based on the ZOC code and the ZOC propagation model. For example, the ZOC 4 denotes a 500 m uncertainty but this uncertainty does not extend into zones with less uncertainty. In addition, depth uncertainty must be subtracted from the depths of soundings and contours.  

Figure 32-3. Snippets from synthetic chart software illustrating two methods to assist mariners reasoning about bathymetric uncertainty while passage planning:  Track Method (left) and Hazard Method (right).

To reduce the cognitive load, we have developed visualization methods that incorporate the relevant calculation and display the results so that safe and unsafe passages can be perceived at a glance.

Two alternative display methods have been designed and for purposes of evaluation the methods have been implemented in CCOMs SimChart (first reported in 2021) and are illustrated in Figure 32-3

In Method 1 (“Track mode”), uncertainty is represented by creating an uncertainty zone around planned ship track using a width equal to the horizontal uncertainty of the underlying CATZOC. Figure 32-3 left illustrates the concept. The planned ship track is highlighted with a transparent yellow band, where the width of the band shows the horizontal uncertainty of bathymetric information in whatever CATZOC regions the track passes through.  In addition, passage is unsafe where (Draft+UKC+Squat+Tide) > (Chart Depth - Bathymetric Uncertainty) and these soundings are shown in bold font and highlighted with circles where they lie in a depth range of concern.

In Method 2 (“Hazard mode”), uncertainty is represented in terms of uncertainty zones around key chart features (Figure 32-3 right). In this method, the hazards are emphasized by drawing transparent hazard zones around them, with the extent of those zones being determined by the horizontal uncertainty. The colored safety zones are drawn both around hazardous sounding and hazardous contours.

Both of the above methods should greatly reduce the cognitive burden on the navigator when reasoning about bathymetric uncertainty and should therefore reduce both errors and time to plan a transit.

Figure 32-4. Snippet from synthetic chart software illustrating the ECDIS method for comparison to the proposed Track and Hazard Methods.

In this reporting period the team focused on fine tuning the two methods and developing an experiment for professional mariners. For comparison, the ECDIS mode has been implemented which replicates the cross-track limits functionality in ECDIS (Figure 32-5). The search distance has been set equal to the average of the horizontal uncertainty of the quality 2 and quality 4, i.e., 275m. This is expected to demonstrate the deficiency of current method as a smaller search distance than the underlying horizontal uncertainty will miss potential hazards for the vessel, whereas a larger search distance will result in false positives which contribute to user’s mental fatigue.

Furthermore, where full coverage bathymetry is available, it is possible to automatically compute hazardous areas taking horizontal and depth uncertainties into account. A method for doing this has been developed and implemented in SimChart (see Figure 32-4). The method used frame buffer color planes, taking into account both CATZOC zones and bathymetry to identify areas that are both at a hazardous depth and within the local horizontal uncertainty. In Figure 32-4 (left), CATZOC zones are shown in red, expanded ship track in green and unsafe bathymetric depths in blue. In combination, this information can be used to reveal potentially hazardous areas (Figure 32-4 right). 

Figure 32-5.  A safety analysis method implemented in SimChart. Left: The graphics frame buffer is used to analyze a potential ship track for hazards taking into account bathymetric uncertainty. Right: The results of analysis with two potential hazards are highlighted in red.

This effort will continue in the next reporting periods with an evaluation study.