This technology gets your car in the right lane — with centimeter-level accuracy

In the not-too-distant future, we can imagine a scenario where autonomous vehicles send signals to each other when an emergency vehicle is coming.

Edge computing in combination with 5G, because of its low latency, is a valuable system to help these vehicles communicate in time to create a clear path for the vehicle, increasing convenience, and ultimately helping lives to be saved.

The same technology can work today to help vehicles communicate with each other, in a process known as V2X, or vehicle-to-everything. However, when Toyota began working with telecommunications company Orange and HERE on a V2X project, they realized there was an immediate problem to solve: how to warn the drivers of today's cars that an emergency vehicle is on its way, and how to make sure they were in the right lane to clear the way.

“We discovered that the embedded global navigation satellite system (GNSS) system in most vehicles can give a position roughly sufficient to know which road it is on, for instance, but not which lane it is in," said Laurent Mussot, a researcher, engineer and project manager at Orange.

The challenge got tougher when the vehicle was in a tunnel, where the signal got weaker, or on highways where there are several lanes to choose from.

“You need to be sure that the position is accurate and reliable," Mussot said. The team put together a pilot based on a combination of HD (high-definition) GNSS Positioning and vehicle dead reckoning (VDR) running for the first time on a remote platform at the edge. The scope was to see if this would achieve the accuracy needed, while checking the feasibility of the system processing onboard sensors in real-time, in the cloud, to send back the corrected position through 5G.

Automatic pilot

A successful demonstration on a test track in France proved that the technology works. The vehicle could be positioned accurately to within centimeters thanks to the sensor data provided by Toyota and to the corrected position from HERE, calculated in the cloud, even when it is in a tunnel. Finally, thanks to the vehicle-to-everything (V2X) platform from Orange, the technology can also warn drivers when other vehicles are braking nearby, helping them to prepare and avoid a collision, no matter what the obstacle.

The findings were demonstrated at the ION GNSS conference in Denver in September.

“The key innovation was that thanks to 5G, edge computing and HERE algorithm efficiency, we have proved that we were able to rely on vehicle data coming in at a high speed and compute the corrected position in the network," Mussot added.

“And it works so accurately, and in real time."

Muriel Desaeger, who is in charge of the incubation of new technologies at Toyota, said the pilot project demonstrates some of the “building blocks" that Toyota will use to further develop its connected and autonomous vehicles.

“In the future, highly connected vehicles will be generating a large amount of data. They will need edge computing so that we don't overload the system, and also for cost reasons," Desaeger said.

Desaeger explained that from the start, the project was undertaken with connected and autonomous vehicles in mind. “This platform will allow vehicles in a specific area to send and receive data so that they know with high accuracy, and within a very short time, where other vehicles are precisely located. When an emergency vehicle arrives they will consider prioritizing it at a lane level to avoid accidents."

The use of cloud architecture for precise positioning capabilities could have significance for the whole automotive industry as automation advances.