Print Friendly, PDF & Email

Unmanned Operations, Virtual and Augmented Reality and Unified Data Models: This year’s COMPIT conference on Computer Applications and Information Technology in the Maritime Industries took place under special Corona conditions – albeit with no lack of expertise, reports Hans Payer

COMPIT 2020 was held in the former monastery Certosa di Pontignano in the Chianti Region of Toscany. Volker Bertram, who[ds_preview] always selects appealing locations and venues for his conferences, made sure that everyone was safe, with very much space in and around the Certosa. This was a very special and intimate meeting with little diversion and good discussions in the arcades or under the cypress trees.

Autonomous ships

The fully autonomous ship – is it a viable vision? Analysis of ship accidents always finds a relatively high number of consequential accidents due to human error. Would the replacement of the human factor by Artificial Intelligence, AI, be a logical consequence? For several years, visionaries have seen the future of AI and autonomous vehicles on the roads, in the skies or on the oceans just around the corner. Only recently have doubts come up about autonomous ships. Can AI cope with the unexpected, calling for quick and reliable reaction?

With an eye on »On-board Human Operators: Liabilities or Assets« Stig Eriksen from Denmark investigates how unmanned operations will influence the ability to detect and prevent near-miss incidents from becoming an accident. He cautioned that in unmanned operations new kinds of incidents might occur. From the analysis of 481 near misses he found that 87% were detected by the crew. On-board safety systems would have detected the incident eventually in many cases, but only after it had become more severe. Monitoring and failure detection devices would have to be significantly extended for unmanned operation.

Realistically, AI cannot fully take over control from human operators today. It is missing intuition and has no »gut feeling«. There are strong indications that – for now – trained and experienced humans will remain necessary as a last line of defence in the automated system.

Marco Bibuli from CNR-INM Genova gave an example for an autonomous water-surface vehicle (ASV). The small catamaran SWAMP (Shallow Water Autonomous Multipurpose Platform) with an elastic outside shell was described, built particularly for the investigation of wetland areas. It is driven by four 360° water-jet thrusters mounted at the fore and aft end of each hull in a way, that they do not protrude below the keel-line. The dimensions are 1,230 x 1,250 x 150 mm. It can be seen as a Water Drone. The vessel is equipped with Artificial Intelligence including self- and imitation learning capabilities and can be trained for its diverse missions.

The project foresees several such units operating together within a network, allowing wider coverage, collect more data and enhance the overall robustness. Bibuli explained the use of AI and machine learning in connection with these vehicles and showed examples, where partly unskilled laymen teach the vehicle to ride a difficult path autonomously.

VR and AR

The use of Virtual Reality (VR) and Augmented Reality (AR) has progressed successfully in the past few years in many fields. Peter D. Petersen from MAN Energy Solutions in Copenhagen delivered insights on »Enhancing the Quality and Reliability of Maintenance Operations using Mixed Reality«. A multimedia application brings interactive and optimised troubleshooting instructions, as well as automatic reporting for two- and four-stroke engines to the maintenance engineer in action, augmenting real life engine components and maintenance scenarios. Using mobile devices or head mounted displays (HMD) the marine engineer has access to visual guidance in 3D, written instructions, diagrams, checklists, videos, spare-parts info and much more. The application has been tested on the ore carrier »Berge K2«, where the system was connected to the planned maintenance system to generate reports and resupply stocks.

Stephan Procee from NHL West-Terschelling reported on an experiment to measure the effect of AR on »Situational Awareness« of seafarers. The interface is based on design principles such as environmental data and ship’s characteristics, particularly regarding manoeuvring. The effect of additional operational information displayed through AR on the situational awareness of the user is observed and analysed.

Participants run two short scenarios on a bridge simulator in which they only observe their situation and respond to a number of probes. Each one starts one of the prescribed scenarios randomly and is also randomly assigned to use AR. The degree of Situational Awareness is determined by the measured reaction time to the individual probes and the related quality of the answers given.

The experiment is still in progress. First results have been presented for 22 professional candidates in the experiment. This is too small a sample for valid generalisations. The questionnaires answered before and after the experiment however clearly showed positive effects. Candidates were surprised by the great potential of using AR for their work as navigators. See-through monitors were preferred to head-mounted monitors HMD.

Axel Friedewald and Robert Rost from Hamburg University of Technology described an AR system supporting the preparation, execution and post-processing of maintenance tasks on board. It includes information on the maintenance history, support for diagnostic analysis, display of maintenance instructions as well as documentation of the work done. Time savings of up to 60% per maintenance task are expected.

In »The effectiveness of VR in Maritime Education« Herbert Kohlmann from MIWB in Terschelling reported on positive experience with VR on a full-mission simulator with big screen. Test results from two groups of students, one with, the other without VR support, clearly showed an improvement of more than 20%.

The VR based training tool »SuSi« (Survey Simulator) developed by DNV GL was described by Volker Bertram. VR in Training is very effective if one does not shy away from considerable efforts. The video game has to be supported by databases with photos of damages, rules and sufficiently complex variations of »virtual worlds« to avoid repetitive scenarios, where the scenario is remembered rather than the general principle.

It contains virtual environments for various ship types and one offshore structure. Technical and safety-related deficiencies can be simulated. Trainees search for these and document them. Practical aspects of creating the virtual environment as well as training options are described and user feed-back is summarised.

»Digital Maritime Training in Covid-19 Times« by Tracy Plowman and Volker Bertram described special aspects in training associated with the 2020 pandemic. Key options for live and and self-paced training, like video conferencing or sophisticated 3D VR – based training were described. Observations associated with learning psychology and lessons learned from the unusual situations like shut-down are discussed.

Unified Data Models

For many years it was accepted that the typical toolkit for ship design contains several independent applications and tools for early, basic and detailed design including several complementary tools for particular design tasks. Transferring and harmonising information between these applications in a robust and convenient way is challenging but offers significant data-integrity and efficiency benefits. Constantin-Alexander Gresens and Carsten Zerbst from Prostep AG described the implementation of such a harmonisation effort. Working with unified data models, they create interfaces between NAPA Steel, SSI ShipConstructor, Aveva Marine and Siemens Teamcenter. This enables yards to work with the harmonised data from initial design to manufacturing preparation without data conversion or disruption.

Nick Danese from NDAR with his generally visionary view, finally described the »Platform of Platforms and Symbiotic Digital Twins«, combining all data bases of a project as well as the diverse digital twins into a universal system – the ultimate goal of highest effciency.

Finally noteworthy are the contributions from Marin Palversa from University of Zagreb on characteristics and design of integrated FEA software (Finite Element Analysis) for ship structural design and Stefan Harries from Potsdam-based Friendship Systems on shape optimisation using Computational Fluid Dynamics (CFD).


Hans Payer