Optimising Data Traffic with the Help of Artificial Intelligence and the Cloud

This article was originally published (in Swedish) by Jarnvagsnyheter.

Developments in globally connected local transport are advancing at a rapid pace, thanks to technological advances such as 5G and artificial intelligence. A company that has not only been on that journey for a long time, but is also the market leader in its niche is Icomera, which has its headquarters in Gothenburg.

A large part of the company’s business model is its proprietary connectivity platform, which is installed in tens of thousands of trains, trams and buses around the world.

“At first, our focus was mainly on how the wireless network could become more stable. Today, we have expanded with everything from real-time travel information to video analytics. Nowadays, our business development revolves principally around how we can create the most value for clients; how our customers can get the best benefits from our connectivity and from deploying multiple applications,” says Mats Karlsson, CTO and co-founder at Icomera.

The technology, which is developed in Sweden, has a variety of tasks; everything from guaranteeing better connectivity for travellers to monitoring equipment on board and analysing and optimising data traffic in real time.

Or as Mats Karlsson explains it: “The trains become like rolling data centres.”

One of the more data-hungry parts of the business is artificial intelligence. The company uses AI for various types of video analytics and monitoring, but also to keep track of service needs and to compare different routes down to the smallest detail. Their solution for monitoring and managing its connectivity platform is a machine-learning system.

“Among other things, we work a lot with anomaly detection. It can be about relatively obvious things like weekends behaving differently than weekdays when it comes to data traffic patterns, but also when it’s about components behaving in an abnormal way. AI is the key to being able to plan optimally after this”, explains Mats Karlsson.

Here, Amazon Web Services (AWS) plays a key role. The company collects a large part of diagnostic data from the vehicles equipped with the technology. It is data that needs to be processed and analysed as quickly and efficiently as possible. The cloud becomes a must when development is going so fast.

“One of the basic philosophies we work according to is to primarily design for the cloud. In the last five years, we have cleaned out older solutions.”