We already have what we need in terms of infrastructure for self-driving cars. N.e.e.d. The problems that remain to be solved are in the code, hardware and their interface.
This is how it works. No wifi required.
Video
Waymo
All you showed was Google AV system. The 5G infrastructure is still missing to take AVs to the level they need to be at. I agree that the problems that remain to be solved are in code, hardware, interface, a
nd also the network. We can argue semantics on what processes are considered critical to driving and what is not. Multiple other car companies and AV companies understand that to ultimately do full self driving right you will need to have it connected to a fast low-latency network like 5G. Intel and Qualcomm would not be building 5G chips designed to be used with AVs if there was no major need.
( I interviewed at BMW research to work on this in 2018 - decided to take a different job. Up to 70K EUR BMW company car after the first year
)
Congrats on your achievement with BMW but you're still out of touch as far as where things are headed. Some critical driving systems will end up on the network. It just makes sense. Companies are eventually goin to offload some of those processes to other servers and/or between vehicles.
Critical driving systems will NOT be on the network. That would be like having a plane's auto-pilot on a network. You would not make a critical system dependent on a non-guaranteed point of failure (the network) when you do not have to.
Having a fully connected vehicle MAY make it safer BUT it is not required for the act of automated driving itself. Part of being fully connected will have nothing to do with act of driving itself - see entertainment, maintenance, security, tracking, diagnostics as per other IOT devices.
Cert gang don't @ me unless you demonstrate that current AV's don't fully work in terms of safe locomotion because they do not have wifi.
Yes some critical systems will end up on the network just like other industries have done in the past.
"
You would not make a critical system dependent on a non-guaranteed point of failure (the network) when you do not have to. "
Do you know what landlines are? Landlines by nature were more reliable than VoIP. Copper lines in the ground. They can withstand power outages. Public Emergency systems and Hospitals have relied upon them for decades. But they're also much more expensive and costlier to keep up. Yet cities, states, countries are moving their 911 emergency system to the network. The entire world is moving to VoIP. Did they "have to"? No. Did we "have to" move to 3G or even 4G? Was it critical? Was it more efficient with increased benefits? Yes. Same thing will happen with other critical systems including those in AVs.