Rolls-Royce has signed a deal with internet giant Google in a move intended to help the British engineering company to develop autonomous ships.
Under the terms of the deal, Rolls-Royce will use Google’s Cloud Machine Learning Engine to further train an AI-based object classification system that it has developed, for detecting, identifying and tracking the objects that a vessel might encounter at sea.
The agreement, which the companies claim is the first of its kind in the marine sector, was signed today at the Google Cloud Summit event in Stockholm. Rolls-Royce has some 4,000 marine customers worldwide, including 70 navies.
Read more: Cargo shipping tech specialist MTI completes blockchain pilot
Maritime machine learning
The Google Cloud Machine Learning Engine uses the same neural net-based machine intelligence software that powers many of Google’s own products, such as image and voice search. It competes against similar cloud-based machine learning platforms from Amazon Web Services (AWS), Microsoft and IBM.
In effect, machine learning methods analyze existing data sets with the objective of learning to recognise patterns, making predictions from previously unseen data. Airbus Defense and Space, for example, uses Google Cloud ML Engine to correct satellite imagery to distinguish between snow and clouds.
According to Karno Tenovuo, senior vice president of ship intelligence at Rolls-Royce, the technology has a key role to play in developing smart ships that pilot themselves, but in the short term, it’s more about recognizing (and hopefully, avoiding) hazards.
“While intelligent awareness systems will help to facilitate an autonomous future, they can benefit maritime businesses right now, making vessels and crews safer and more efficient. By working with Google Cloud, we can make these systems better faster, saving lives,” he said.
Read more: Dubai shipping company Topaz links ships to shore with Maritime Connect
Shipping forecast
The intention is for Rolls-Royce to use Google Cloud’s software to create bespoke machine learning models that can interpret the large and diverse marine data sets that the engineering company has created. This data must be relevant and present in sufficient quantities to get statistically significant results from machine learning and the resulting models will be evaluated by using them in practical marine applications, so that they can be refined over time.
In the longer term, Rolls-Royce and Google say they intend to undertake joint research on a range of areas: unsupervised and multimodal learning; the use of speech recognition and synthesis in marine applications; and using machine learning on board ships.
The goal of this research, of course, is smarter ships: safer, easier and more efficient to operate for crew that have a better understanding of the environment in which a vessel finds itself. This will be achieved by combining data from a on-board sensors, existing ship systems such as radar and other IT systems, such as marine databases and mapping applications, for example.