Where It Comes Down to Hundredths of a Second in Extreme Cases

Edge Computing in the Digital Mobility

By Gabriele Strobel, Software AG

What will mobility look like in the future? Will we still need drivers? Or will our vehicles drive themselves entirely, with the driver only providing the destination?

While manufacturers are developing technologies at top speed for networked automated driving and are already testing initial prototypes, the development framework and conditions are still lagging behind somewhat. Autonomous mobility can only be realized with an intelligent infrastructure that reliably provides its service while remaining unobtrusive in the background. And Edge Computing will play an essential role in the Digital Mobility.

The goal of the research project Cartox2, supported by Germany’s Federal Ministry of Transport, is to develop a service platform for base services for networked and automated driving. This includes information on mobile communications coverage, the availability of edge computing as well as the analysis of traffic risks. Along with digital maps, these basic services are essential for anonymous driving. Thus, alternative processes need to be implemented for routes where car2car communication is limited, such as in urban canyons. Cartox2 focuses not only on the traditional auto industry with its work, it also addresses providers of digital mobility services who are actually outside the industry.

When a car is driving autonomously, a signal disruption or information gap could cause extreme safety problems, because at some point down the road the technology will require traffic to be controlled by an intelligent infrastructure. For example, public and individual transportation will need to be aligned, as will the switching for traffic signals.

Cartox2 is making use of a key IoT technology with edge computing, meaning decentralized data processing—which is even handled in the sensor itself in extreme cases. Cloud computing reaches its limits with autonomous driving because of its relatively long latencies. For physical reasons alone, it would take too long to send data from a car to the data center, process it there and then send it back to the car. Hundredths of seconds count in autonomous driving, which is why data must be processed locally—on the edge–in the nearest cell tower, for example.

Cartox2 uses many data sources including open data such as the broadband atlas, the DB network radar and weather data. Test cars also collect their own connectivity data. The Cartox2 pilot project is being tested in the Dresden metropolitan area. Its topography and morphology is representative for all of Germany so that predictions for the entire country can be derived with the help of statistical methods.

The project partners are:

  • Fraunhofer Institute for Transportation and Infrastructure Systems (consortium leader)
  • hrd.consulting
  • TU Dresden, Faculty of Transport and Traffic Sciences
  • MechLab Engineering UG
  • Software AG

The project launched on July 1, 2017 and will run through June 30, 2020.