I have always presented the position that, to have fully autonomous (level 5) vehicles, there must be two-way communications between the vehicle and the infrastructure (V2X). To be clear, the “X” means everything, fixed and mobile, from people to baby carriages, even pets.
The way I envision this happening is to put sensors with radios everywhere, including into the road base. However, putting them into the road base is a daunting task fraught with challenges. One that will take decades for a ubiquitous solution.
So why not look for a more workable solution that can be deployed in a much shorter time span and will work as well as embedded sensors in many circumstances. As well, this is an opportunity for 5G to shine.
Now comes a study by Bosch – Mercedes, TomTom, Osram, Nokia, and a couple of universities – that has given validation to two-way communications as a significant element of V2X – the “birds-eye,” view of traffic as the study calls it.
Hear-hear. This may well be the first step towards level 5 autonomous vehicles. While such a communications system will level up such vehicles, it still is not the ultimate in V2X communications on a ubiquitous scale.
However, this will go a long way towards that lofty goal and could, potentially, make fully autonomous vehicles possible in places where the street lighting is prolific.
The reason this will be successful is simple. It will remove the condition that all data be captured by the vehicle and decisions made are based upon that data.
While sensor technology has come a long way, it is far from foolproof. Hence, all the guarded conditions around what limited autonomous deployments are currently in existence.
Finally, some validation has been given to the position of two-way, V2X and back again, communications in this segment.
A multi-year research project in Germany has come to the conclusion that to have fully vetted autonomous vehicles will require a combination of technologies, including Internet of Anything/Everything (IoX) sensors, edge computing and 5G connectivity. And, street lighting is the key ingredient, sans embedded roadway sensors.
The trials were focused on city infrastructures. That only makes sense since there are not a lot of streetlights out in the boonies. However, if level 5 autonomous vehicles are needed anywhere, it is in cities where the biggest traffic management challenges occur.
The project site was the city of Ulm. The undertaking combined cellular LTE and 5G radios, video and lidar sensors and multi-access edge compute (MEC) hardware mounted on the poles reaching up to six meters in height. It lasted three years.
What makes this both different, and sound, is that this is the first scenario where two-way communications were used in real-time. Such data is what the industry has been sorely lacking and why autonomous vehicle progress has hit a bit of a speed bump. And, while sensor technology has come a long way, it can only do so much.
The fact is that this vector adds, at least, an order of magnitude of capabilities over vehicle-height sensors. Such an overview capability can consider a multitude of difficult, or impossible, to see objects such as vehicles obscured by larger vehicles, cyclists, pedestrians – anything difficult to spot at eye level.
It can also reveal hidden and distant objects that can come into a vehicle’s path long before the vehicle sensors can spot it. Such elevated video, radar, and lidar provide a 360-degree field of vision. Certain locations can also see distance (such as poles on corners) that vehicles cannot see until they are at the intersection.
For now, this trial uses application-specific hardware and software that acquire and process the images and signals from the streetlight sensors and transmit the data to vehicles. Currently, the medium is both LTE and early 5G connections. However, once of vehicle data analysis systems become more sophisticated, the latency of 4G will preclude it from being used. 4G will also be challenged with the sheer volume of traffic. Therefore, 5G will be the platform that will support all of this, and at much higher frequencies. Current frequencies under consideration are 28, 37, and 39 GHz of the licensed band and 64-71 GHz of the unlicensed bands. However, exactly where this will play is still muddy at this point.
At this time, because 5G is not available, the designers tweaked LTE to achieve near 5G performance. Of course, this is not practical in the final iteration but, it does paint a portrait of its possibilities. Also, what will play a large part in this is the MEC. Once that has evolved, combining it with 5G will provide the bandwidth and latency requirements to handle the immense volume of data, in real-time.
However, as usual there is a spoiler. In this case it is the ability to have access to streetlights. The challenge here is to work with municipalities since such poles, with few exceptions, belong to the city, or the utility company.
And, we all know that dealing with these entities is a piece of cake, right?