The data revolution in shipping

in Trends by

Artificial intelligence in today’s world:

AI or Artificial Intelligence is a blanket term that refers to many computer systems having “intelligence” in some form or another; even if the program is highly overseen by humans. Nevertheless, “Narrow AI’s” (AI’s that do not have general autonomy), while far less interesting constitute far more of what the society today perceives as AI.

The strength of an AI lies in recognizing the underlying patterns and drawing inferences from data that is too complex for human analytical capabilities and finding solutions to problems accordingly. These can be invaluable in diagnosing problems and finding novel solutions not only in businesses but in a vast array of fields ranging from the mundane traffic flow problems to complex neuroscience problems. A global McKinsey survey states that there is a 25% year over year increase in the adoption of AI in businesses.

AI in the shipping industry:

Logistics is a field that can benefit greatly from AI, it is pretty intuitive to grasp that the raw data, processed correctly will point out logistical bottlenecks. It can also be used to run simulations and test solutions to streamline said bottlenecks. The maritime industry values optimization and therefore is fairly free of obvious hindrances but AI can go the extra mile and make improvements by fine-tuning operations such as container routing, refueling, etc. to the benefit of all.

Completely autonomous ships are the first thing that comes to mind when thinking of AI in the maritime industry. The first fully autonomous vessel was launched by Rolls Royce and demonstrated in 2018. The Mayflower Autonomous Ship (MAS), an autonomous trimaran research vessel embarked on its first journey on September 15, 2020. It is a vessel produced for studying critical issues such as global warming, micro plastic pollution, and marine mammal conservation by the joint efforts of IBM and ProMare and others.

Boston based startup Sea machine robotics builds remote command vessels and completely autonomous vessels for commercial purposes. They also have a workboat system through which older vessels may also be retrofitted for less than 100,000 USD. Sea machine recently announced a contract with Maersk to use situational awareness technology that will be applied to their ice-class vessels. Sea machines expect this to reduce operational costs by 40 percent and increase ship productivity by 200 percent.

Silo.AI and Awake.AI, two private AI-based companies, are collaborating to improve the situational awareness of port logistics. Their programs have yielded better predictions of arrival times of ships, an improvement of over 80 percent. They also have succeeded in automating the challenging manual analysis of ship routes. This expedites cargo logistics planning and also enhances the detection ability of potential exceptions that may arise en route so that solutions can be applied proactively. Applying AI to shipping ports also helps in more efficient automation; by predicting the arrival of cargo and thereby mobilizing other vehicles such as dredge trucks and forklifts such that the idle time is minimized.

The challenges to overcome:

As impenetrable and invincible today’s media may make AI seem, it is not without its fallacies. An AI is not a magic eight ball that has all the answers. When faced with complex problems, it takes consummate skill to arrive at the solutions to some and others may not be solvable at all.

Many experts argue that the most crucial part of most AI’s is the available dataset. Lack of good source data can completely hamstring an AI. Lack of sensory data aboard most ships and a reluctance to share data based on having a competitive advantage can hamper the growth of the field. Another problem is inaccurate or erroneous data. When collected across the whole supply chain, there is a significant probability of errors creeping into the input data, especially when the data is being collected by workers not intimate with the intricacies of an AI.

Lack of expert knowledge is another great disadvantage. In most Narrow AI’s, building a framework that can properly process and analyze data is a skill that is in sadly short supply. As much as the field of AI has been blown up many companies lament the lack of skilled workers in the field.

The looming dread of AI’s taking over jobs may not be completely unfounded. As the supply chain and the logistics industry is “optimized” there is a reasonable probability that a few jobs will become obsolete. A good example is that 30% of the cost of a ship’s journey is crew compensation, shipping companies stand to make a tidy profit by switching over to completely autonomous ships. 

The rewards of AI integration:

Like any other field, the shipping industry stands to gain a lot from the data revolution. Charting optimized routes to save fuel and avoid obstacles, controlling port and ship equipment, and forecasting a ship’s degradation is only the tip of the iceberg. Hitachi Europe Ltd. Partnering with Stena Line, one of the largest shipping companies in Europe, implementing AI on ships to reduce fuel consumption for the green initiative.

In June 2017, the shipping giant Maersk Co. was the victim of a cyberattack that caused losses of over 300 million dollars for the company. Cybersecurity threats are usually preceded by unusual network activity. An AI can go through the network traffic and detect network anomalies that are too subtle to be noticed by humans.

The improvement of safety is another aspect that cannot be glossed over. 75% of marine accidents are caused by the human component of the maritime industry (Allianz 2019). So an increasing dependence on AI not only to alleviate the burden of the crew but also to detect weather patterns, crime hotspots, technical anomalies, port conditions, etc. and make an informed decision that is to the benefit of all. 

Among the myriad of challenging and complex problems that are solved by the application of AI, there are also the mundane aspects such as the processing of paperwork. The maritime industry is infamous for the vast amount of paperwork that must be meticulously filed every day. While this is a daunting task for people, computers are especially suited to such tasks and can do the work without errors much more economically.

Huge inefficiencies are inherently present in the global supply chain. AI will inevitably find its way to more and more aspects of the shipping industry as time goes on. Nautic expo states that the freight revenue will reach 205 billion dollars in 2023 from 166 billion dollars in 2017 owing to the widespread adoption of AI.

FleetMon and AI technology:

“The correct prediction of subsequential port-to-port routes plays an integral part in maritime logistics and is therefore essential for many further tasks like accurate predictions of the estimated time of arrival.”

(D. Marten, C. Hilgenfeld, A. Heuer 2020, p. 2)

FleetMon works on a scalable AI-based approach to predict upcoming port destinations from vessels based on historical AIS data. The respective method is mainly intended as a fill in for cases where the AIS destination entry of a vessel is not interpretable. We succeeded to find a stable and efficient in-database AI solution built on Markov models suited for massively parallel prediction tasks with high accuracy. The research is part of a funded project called PRESEA (“Real-time based maritime traffic forecast”). The PRESEA project is funded by the German Federal Ministry for Economic Affairs and Energy (BMWi). Project management organisation is administrated by the Project Management Jülich (PtJ) within the framework of the call “Real-time technologies for maritime security”. The project is running from June 2019 until November 2021.

If you’re interested to know more about this AI-based approach, please read the open-source paper published by two of our data experts:

D. Marten, C. Hilgenfeld, A. Heuer: “Scalable In-Database Machine Learning for the Prediction of Port-to-Port-Routes”, Open Journal for Mobility and Transport, pp. 2-10, 2020.