Triton 2025 - Magazine - Page 42
PAGE HEADER
PERSPECTIVES
“The complexity of the maritime environment with its dynamic
conditions, rare failures, and a wide range of vessel types,
makes AI modelling both challenging and rewarding.”
Unlike previous models, which often
treated ships as predictable dots on a
chart, STAPS takes into account the
real-world imperfections of maritime
navigation. It recognises that most
allisions result from human error something often acknowledged but not
regularly addressed in design codes.
Toward Practical
Applications and RealTime Monitoring
While Axel’s current work focuses on
informing bridge design, building structures
to withstand probable rather than
hypothetical impacts, the implications of AI
extend far beyond construction.
Imagine a real-time system using
AIS and AI to monitor ships as they
approach a bridge. Such a system
could issue a warning: “If you don’t
change course now, you will hit the
bridge.” This predictive alert could
be transmitted to the ship’s master
or a shore-based control centre,
potentially averting disaster with
seconds to spare.
This kind of proactive safety
intervention marks a shift from
passive risk management (building
stronger structures) to active
prevention (modifying vessel
behaviour). The possibilities are
especially relevant as global shipping
routes grow more congested, and
42 / Triton 2025
infrastructure becomes more complex
and expensive.
Collaborative Efforts and
Industry Impact
Axel’s collaboration with The Swedish
Club (TSC) is a promising example of
how academic research and industry
data can complement each other. While
Axel uses AI to detect failures from
ship behaviour patterns, TSC offers
validation through real-world data on
insurance claims and near-miss reports.
This partnership may help to con昀椀rm
whether anomalies spotted by AI
actually correlate with known incidents,
strengthening the model’s accuracy.
He also sees potential in working
more closely with insurers and
maritime operators to build a more
comprehensive dataset of failure
types, frequency, and severity. This
could eventually inform not just bridge
design, but also training programmes,
operational protocols, and insurance
underwriting practices.
What About the Ships?
Despite all the insights, Axel is clear that
operational change on vessels is not the
immediate focus of his research. “Ships
often know what to do - they just don’t
always do it,” he says. Whether due to
fatigue, distraction or misjudgement,
human error continues to be a dominant
factor in maritime accidents.
That said, integrating AI-driven insights
into bridge management, traf昀椀c control,
and vessel monitoring systems could
drive behavioural change that could
lead to greater safety.
Understanding how these changes
impact operational costs, compliance,
and safety records will be a crucial area
of future research.
Charting a Safer Future
Allision prevention requires a combination
of engineering, data science, and human
judgment. By simulating real-world vessel
behaviour and identifying the most probable
failure scenarios, AI offers a revolutionary
tool for designing safer maritime
infrastructure and minimising risk.
As Axel adds, “The complexity of the
maritime environment with its dynamic
conditions, rare failures, and a wide range
of vessel types, makes AI modelling both
challenging and rewarding.”
Through models like STAPS and
collaborations with insurers and
operators, researchers like Axel are
pioneering a smarter, safer future
for global maritime navigation. As
data continues to accumulate and AI
becomes more adept at interpreting it,
the dream of a predictive, preventive,
and resilient maritime safety system is
not far from reality.
