In the future, we will be able to work while commuting by car, on a route without traffic. The personal benefits from the saved time and environmental impact from the carbon emissions avoided are attractive. And this is only one example of what it will be like to have self-driving cars on the road.

Autonomous driving offers many promises for the future, and the deployment of 5G will be crucial to enable safety protocols.

A road is a highly dynamic environment, with conditions that change by the second. Other vehicles, bikes and pedestrians all create a situation that a driver needs to assess and make decisions about. An autonomous vehicle needs to do the same, and what enables this process for the car is data about its environment. Data that is extremely precise, that when pieced together creates the car’s perception.

Broaden the perspective

In early versions of autonomous vehicles, information about the environment around a car was collected using sensors on the car itself. Then, the data the sensors collected was processed by a computer on board. This approach was, and still is, prohibitively expensive because it requires so much onboard technology. The price point would be more similar to a small apartment, not an accessible car designed for mass-market production. The car was also limited to the data that onboard sensors could collect within their own range, from the car outwards.

That is why we are now looking at how to make that perspective so much broader, while also limiting onboard technology, to work towards a safe vehicle with a realistic market price point. 5G is a foundational element to achieve this.

As a car moves through its environment, connecting it to other vehicles and infrastructure will give the car a significantly increased amount of data and computing power to work with. Other vehicles that are ahead can warn of immediate dangers on the route, and sensors that are stationary can collect different perspectives, altogether allowing a certain amount of predictive modelling. This massive amount of data coming from the sensors in the environment can then be processed in computers close by and sent back to the car as information it can make decisions about. It is a collaborative approach to mobility, instead of an individual one.

With the increased amount of information from the collaborative approach, a car can now see a lot more, and deliver a safer ­driving experience.

Raphaël Frank, assistant professor and senior research scientist, University of Luxembourg

Deliver a safer driving experience

These increased capabilities that collaborative sensing gives an autonomous vehicle are best compared to the difference of a human driving in a rainstorm or a sunny day. In the earlier versions of autonomy, a car was all alone, receiving just enough data from its sensors to reach its destination, but still has an obscured view that wouldn’t allow it to react as quickly to changing road conditions. With the increased amount of information from the collaborative approach, a car can now see a lot more, and deliver a safer ­driving experience for both the passengers and ­other actors in the environment.

A road is a dynamic place though, so all the data needs to be exchanged extremely quickly. The information the sensors collect needs to be sent to the nearby computer for processing and sent back to the cars in time for the cars to make real-time decisions. That is why 5G is crucial as collaborative approaches have already been tested with 4G and the network could not deliver the information fast enough for it to be useful for the vehicle.

5G, and the infrastructure that will be rolled out to deploy it, is predicted to deliver the low latency required for collaborative autonomous driving. And tests are already planned, one site existing in the south-east of Luxembourg as part of the 5GCroCo (Fifth Generation Cross-border Control) project being run by the 5G Infrastructure Public Private Partnership (5G PPP). There, they are looking at the impact of deploying the cooperative sensor technology in a cross-­border scenario.

Making a driverless commute a reality

That is only one example of the support ­Luxembourg offers to autonomous driving, with the work we are doing at SnT (Interdisciplinary Centre for Security, Reliability and Trust) with the 360Lab looking at the ­interdisciplinary questions that arise from autonomous mobility, from AI in the car to social responses to the car. It is an exciting field, and the positive research environment in Luxembourg is beginning to attract more interest, including Civil Maps, a company supplying 3D mapping for autonomous ­driving that recently set up their European headquarters here.

We have a long road ahead of us to make the scenario of a driverless commute a reality, and there are multiple factors to achieve it. 5G is only one of these but is crucial to ensure the safety of autonomous vehicles, tying the future of this promising technology to the development of that network.

Cet article a été rédigé pour l’édition magazine de   qui est parue le 25 novembre 2020.

Le contenu du magazine est produit en exclusivité pour le magazine, il est publié sur le site pour contribuer aux archives complètes de Paperjam.

Votre entreprise est membre du Paperjam Club? Vous pouvez réclamer un abonnement à votre nom. Dites-le-nous via