Autonomy develops in stages: assistance, partially automated functions, supervised processes.
Port logistics, remote monitoring, defined coastal profiles and recurring assistance functions.
For operators, intermediate solutions are often more relevant.
Start from the function, not from the vision.
The IMO distinguishes four degrees of autonomy for shipping, from degree 1 (crew on board, decision support automated) to degree 4 (fully autonomous operation). In practice, the industry operates almost exclusively at degrees 1 and 2. Degrees 3 and 4 exist in pilot projects and on short, controlled routes – not in liner services or tramp shipping.
At level 1, autonomous functions are already widespread, even if they are rarely labelled as such. Dynamic positioning on offshore supply vessels, automatic ballast water systems with sensor-based control, or main engine monitoring with automatic alerting – all of these are assistive autonomous functions. They support the crew but do not replace it.
Level 2 – remotely monitored operations with reduced crew – is found in coastal shipping and among harbour tugs. The Norwegian Yara Birkeland is the most prominent example: an autonomously operating short-range container carrier between port and loading point. Yet even this project operates with extensive remote monitoring and strict operational boundaries.
Technically, autonomous systems rely on a combination of sensors (radar, lidar, cameras, AIS), situational modelling (detection of other vessels, obstacles, weather effects) and decision algorithms. The greatest technical hurdle is not the hardware but the reliability of situational modelling under real conditions: sea state, restricted visibility, unpredictable behaviour of other traffic participants.
In the port domain, autonomous applications are further advanced. Automated straddle carriers, driverless transport vehicles and autonomous cranes are already in operational use at large container terminals. The decisive difference: these systems operate in controlled environments with defined routes and predictable traffic.
The pragmatic view is directed not at fully autonomous vessels but at the assistive functions that are available and economically sensible today. Three areas stand out.
First: condition-based monitoring with automatic anomaly detection. Systems that continuously evaluate vibration, temperature and pressure data and trigger automatic alerts when values deviate from the normal range. This reduces unplanned downtime and enables predictive maintenance planning.
Second: voyage optimisation. Algorithms that combine weather, current, draught and schedule data to provide route recommendations that reduce fuel consumption. The decision remains with the master, but the data foundation is significantly better than with manual planning.
Third: automated documentation and compliance. Systems that automatically convert operational data into the prescribed formats – for instance, for CII reporting or EU ETS data capture. This saves manual effort and reduces error sources.
All three areas share a common prerequisite: they depend on solid data quality and clear processes. Without these foundations, even the best autonomous function does not work reliably.
Investment in maritime autonomy is unevenly distributed. Scandinavia leads in pilot projects – driven by short coastal routes, government funding and a historically strong maritime technology base. The Yara Birkeland in Norway and Finland's ferry projects are concrete examples.
In Asia, shipyards and classification societies are investing. Samsung Heavy Industries, Hyundai and ClassNK are working on autonomous navigation systems for newbuilds. The focus here is more on integration into conventional vessel designs than on radically new concepts.
In Europe, port operators focus on automation: autonomous terminal vehicles, automated crane operation and intelligent traffic management. These investments are commercially driven – they reduce labour costs and increase throughput. The autonomy label is secondary; the issue is efficiency.
For an operator who must evaluate autonomy offerings, four test questions apply: 1) What specific problem does the autonomous function solve? If the answer remains vague, that is a warning sign. 2) What data is required, and is it available? 3) Who monitors the system during operation, and what happens in the event of a failure? 4) How is the function treated regulatorily – by class, flag state and PSC?
The honest answer for most operators is: investments in assistive functions at level 1 are worthwhile today. Investments in level 3 or 4 are irrelevant for daily operations and will remain so for the coming years. This is not pessimism but engineering judgement.
When a vendor pitches an autonomous or semi-autonomous system, the sales material tends to blur the distinction between what the system does today and what a future software release might do. A useful test is to ask for a live demonstration on the operator's own vessel type and route profile rather than a reference case from a different trade. Ask specifically what happens when a sensor fails, when GPS signal is degraded, or when the vessel enters a traffic situation the training data did not anticipate. Vendors who answer with concrete fallback procedures are describing a mature product; vendors who answer mainly with reassurance about ongoing improvement are describing a beta.
It also helps to ask who carries liability if the assistive function gives wrong guidance and the crew follows it. The answer often reveals how confident the vendor actually is in the system's reliability, and whether the contract reflects that confidence or simply shifts the risk back onto the operator.
An assistive system only delivers value if the crew trusts it enough to use it and stays skilled enough to override it. Both conditions require deliberate change management, not a one-off familiarisation session before delivery. Crews need scenario-based training that includes deliberately degraded conditions, so officers learn what the system looks like when it is wrong, not only when it works.
Equally important is a clear operating procedure stating when the officer of the watch must fall back to fully manual operation and who signs off on that decision. Operators who skip this step often see two failure patterns: either the crew ignores the system's output because nobody explained how to interpret it, or the crew becomes over-reliant and loses the manual skills the system was meant to support rather than replace.
Free initial consultation – we analyze your situation and find the best path forward.
Request Consulting