Apr 23, 2019 · Digital innovation

It’s called ‘artificial,’ but the intelligence is real

It has a voracious appetite and needs to be fed constantly to work at top efficiency. The more it takes in, the smarter it gets. And the smarter it gets, the safer and more efficient Royal Caribbean ships become.

Artificial intelligence, or AI – also known as machine learning – both imitates and, in at least one aspect, surpasses human thinking to forecast possible outcomes of combined data. It does this by seeing patterns in the data, including anomalies from what is otherwise expected, much better than the human brain.

And now AI is integral to optimally routing RCL ships for weather safety, fuel efficiency and guest comfort. On the newer ships, it can even keep track of how much energy a single oven is using in the ship’s galley.

As Anshul Tuteja, RCL director of energy management, fleet optimization for global marine operations, explains:

“We have a multiple sensor network on board. It’s a neural network, like in your brain cells. It brings (raw data) to a consolidated machinery automation system station, and from that station we collect all the operational data tags.”

By imitating the brain’s highly complex neural network, the digital machinery learns to think and process the data input, spotting patterns and making predictions.

“Just to put things in perspective,” Tuteja adds, “as of today, we collect roughly two billion data tags from our fleet yearly. There could be 2,000 different variables that we collect from a ship daily.”

The main task is to optimize a ship’s operation in real time and to predict what’s ahead. Spotting anomalies is a relatively new addition to Royal Caribbean’s AI functions.

Once the system has the raw data, it creates an optimal performance model against which real-time ship operations are measured.

“If the actual measurements are not matching the simulated model or a simulated forecast, then the system issues an anomaly,” Tuteja continues. “For example, if suddenly there is leakage from an engine, you can be pre-advised if a part is supposed to fail or if the engine is not going to perform.

“Those kinds anomalies you can detect in advance and then you can raise an alarm.”

While anomaly detection is still in the experimental stage at Royal Caribbean, Tuteja says he believes the fact that data is auto logged and continually fed into the ever-growing cache used by AI puts RCL ahead of the rest of the maritime industry.

“All of this is kind of making sure that we are also digitally transforming ourselves,” he says. “That’s the journey for us, that we have all the data, all the business intelligence, all available on digital platform so that any captain, any chief engineer, any shoreside personnel can just open a smartwatch or smart tablet or a smart computer anywhere and see what exactly is happening and what could happen if they don’t take an action.”

While Tuteja stresses that ship-handling decisions are always the purview of the captain, the confidence level of the data system continually improves, offering viable options to each master.