With the agreement signed between Rolls-Royce and Google, it is aimed to develop smart awareness systems that will increase the safety of existing ships and enable autonomous ships to come to life.
This agreement, a first in the maritime industry, was signed at the Google Cloud Summit in Sweden on October 3, 2017. In line with the agreement, Rolls-Royce aims to further train the company’s artificial intelligence-based object classification system using the Google Cloud Machine Learning Engine, thereby increasing the ships’ ability to detect, identify and track objects they encounter while cruising.
Karno Tenovuo, VP of Smart Ships at Rolls-Royce, said: “Intelligent awareness systems are a development that can enable ships and their crews to operate more efficiently and safely, as well as enabling autonomous ships to come to life in the future. By collaborating with Google Cloud, we can make these systems run faster and better and save lives.”
Eva Fors, Sales Manager, Google Cloud Scandinavia, said: “Exploring the possibilities offered by machine learning, Rolls-Royce is in a position to bring serious innovations to the industry by combining the latest technological developments with its vast knowledge of the maritime industry.” said.
The Google Cloud Machine Learning Engine uses the network-based smart machine software that underlies many of Google’s products, such as image and audio search. Machine Learning consists of a set of algorithms, tools and techniques that mimic human learning to solve specific problems. These methods detect patterns or repetitive phenomena by analyzing existing datasets, and make predictions about data that they have not seen before, using the information they gain from it. The larger the dataset, the more complex the patterns the model can detect and the more accurate its predictions. Well-trained machine learning models today can perform predictive analytics faster and more accurately than humans.
Using Google Cloud software, Rolls-Royce will develop custom machine learning models that can interpret the vast and diverse datasets created by it for shipping. Rolls-Royce’s expertise in the maritime industry will also play a role in the preparation of data to be used in training models, ensuring that it is statistically significant and closely related to real applications. As part of the machine learning process, the predictions of the models will find practical application at sea, enabling the models to be further developed.
The fact that the software is accessible via the cloud means that models can be developed anywhere in the world and thousands of users have worldwide access. Therefore, it is possible to train models on data large enough to be measured in terabytes. This will become even more important as autonomous ships become more common.
Rolls-Royce and Google plan to conduct joint research on unsupervised and multimodal learning in the long term. The companies will also test whether their speech recognition and synthesis applications are viable solutions for the human-machine interface in the maritime industry. The two companies are also aiming to improve the performance of LAN-based computing on ships through the use of open source smart machine software libraries such as Google’s TensorFlow.
Intelligent awareness systems will enable ship crews to better understand what is happening around them, allowing ships to be operated more safely, easily and efficiently. For this, data from a series of sensors will be combined with information provided by systems such as the Automatic Detection System and radar already on board. Data from other sources, such as the global database, will also play a role.
