I've heard that we will not be able to have autonomous cars until 5G is everywhere. 4G is simply not fast enough to support any near real-time car to car communication.
This is kind of true, but only from a certain angle.
For one, autonomous vehicles (the ones most people care about anyway - Waymo et al) are not widely using connectivity between vehicles - who wants an autonomous vehicle that will only stop for other compatible vehicles? Instead they are almost entirely reliant on visual sensors etc.
What has changed between 4G and 5G is the device to device protocol (which is not widely used and is completely incompatible with normal mobile signals) has been improved to support vehicles travelling at high speeds. But the thing is, there was already a protocol to enable that kind of communication and 5G does little if anything to improve on it.
Autonomous vehicles are probably never going to rely on network connectivity to work properly - the risk of losing your connectivity is too great.
> the risk of losing your connectivity is too great.
Additionally the risk of benign or malicious interference is too great. We can't live in a world where a 14-year old with a soldering iron and $20 in parts can jam 12 lanes of traffic at rush hour. Or where a solar wind or foreign enemy can cause mass casualty events such as disconnecting all autonomous vehicles simultaneously.
I wonder if a low bandwidth directional link might be best. Where all cars in a "pack" agree they have the same firmware. Then do each others calculations and then confirm whether they agree by only checking a hash of the calculation. This would be a much smaller payload and could be transmitted on redundant channels on different bands. It would at least be harder to jam.
> where a 14-year old with a soldering iron and $20 in parts can jam 12 lanes of traffic
Well... not to be argumentative or anything, but a 14-year-old with a smaller budget could easily cause similar chaos on a busy highway by, say, rolling a heavy object down a hill onto it or throwing some spiked gardening tools onto the road at night. Although they don't generally do that, protesters do pretty much the same thing (that's how we all first started hearing about BLM, recall).
> but a 14-year-old with a smaller budget could easily cause similar chaos on a busy highway by, say, rolling a heavy object down a hill onto it or throwing some spiked gardening tools onto the road at night.
Doing that puts them at severe risk of detection and capture, doesn't scale well, is pretty easy to remedy, and dealing with foreign objects on roadways is pretty easy (people drive around them, and maintenance personnel and/or police move them off the road) and needs to be because it happens all the time without any malicious actors involved.
Yup, it's way too easy to dump RFI that throws off everything.
Without fail my 2 meter radio keys up in about 70% of the drivethroughs I go through. Some of them are probably due to overloading the front end, some of them probably aren't filtered well enough.
Had a microwave that would take out WiFi in a ~200ft radius when it was running.
True autonomous cars need to work without any real time car to car communication. After all other cars on the road won’t be using communication protocols as they are driven by humans, etc.
True that they don't need it to work. But imagine the efficiency and safety rewards if they do have it.
You could do efficient ingress and egress routing if adjacent cars had the means to communicate their destinations, rather than using 100-year-old technology like blinking lights.
'If intelligent traffic management becomes a reality in future cities, more and more self-driving cars will appear on the road. To guarantee traffic safety, when a control command, braking for example, is sent to a car, the car must receive the command within 1 ms.
The latency of a 4G network cannot meet this requirement. With the latency of 4G network, a car driving at 100 km/h still moves 1.4 m from the time it finds a obstacle to the time when the braking command is executed.
Under the same condition, with the latency on a 5G network, the car will move just 2.8 cm, and this performance is comparable with the standard of an anti-lock braking system (ABS).'
> To guarantee traffic safety, when a control command, braking for example, is sent to a car, the car must receive the command within 1 ms.
> With the latency of 4G network, a car driving at 100 km/h still moves 1.4 m from the time it finds a obstacle to the time when the braking command is executed.
I'm a fan of self-driving cars, but I would never buy one that depends on an external network to know when to brake.
Full disclosure: I work for Veniam on connected vehicle technology, including LTE, Wi-Fi, and DSRC stacks.
Unfortunately there's also still no agreement on what 5G will actually be, what technologies it will entail, or how it will be deployed. Of course, since none of this is defined yet, there are also no guarantees on what performance will be achievable in practice.
The industry will need to solve this issue long before 5G is widespread. The solution will likely involve DSRC, C-V2X, or LTE-V/LTE Direct, which are tailored for the vehicular communication use case (involving both vehicle-to-vehicle [V2V] and vehicle-to-infrastructure [V2I]). It's unlikely a single technology will be able to handle all the challenges of a V2X scenario (highly mobile nodes, congestion, low SNR, requirements for sub-millisecond connection times), and so a second connection management layer will have to be built on top of multiple existing PHYs to enable simultaneous low-latency, high-throughput, and secure communications.
(Note that the linked article is actually about in-vehicle communication, however.)
Fun facts: ITU-T criteria for 4G is 1 Gbit/s nominal data rate. So LTE is/was 3G. LTE-Advanced, which has seen deployments in 2017/2018, can get to 1 GBit/s on paper. Android has a "LTE+" indicator for LTE-Advanced.
5G marketing says this a lot, but operational hard dependecy on always available low latency 5G communication is clearly fundamentally incompatible with the safety requirements.
Consider:
> “The big difference with 5G is that when you start to talk about “autonomy” and factories, cars and hospitals thinking for themselves, they will rely on split-second connectivity to do so – with no room for error,” said Aicha Evans is senior vice president and general manager of the Communication and Devices Group at Intel Corporation, in February just ahead of Mobile World Congress.
Cars don't need a fixed hub to communicate, they can talk vehicle-to-vehicle because they are mostly subject to local effects. See "v2v" research on this topic.
So car-to-car communication is likely to be supported by industry before mobile-to-mobile communication is widespread. Based on cars requiring real-time, local data. Sad.
The only mobile to mobile communications that make sense are widespread - WiFi direct and Bluetooth; both AirDrop and Airplay utilize WiFi direct. Peer to peer protocols on mobile is a great way to achieve horrible bandwidth, even worse latency and excellent battery drain. There are no benefits vs just using direct connections over ipv6 (where available). If you need local real-time data on a smartphone, just use a BLE beacon for low bandwidth or WiFi multicast for high.
Also V2V is vaporware that isn’t actually needed for safe self-driving cars, because the cars have to safely interact with human-driven vehicles (the majority of cars on the road for the next 10 years).
Modern radios’ high power efficiencies are achieved through multiple means; adaptive signal strength and low duty cycles are some of those.
Both antennas being small and low on the ground means much lower RSSI, which causes higher adaptive transmission power and higher duty cycle due to lower SnR meaning lower bandwidth and increased retransmissions. Normally, device negitiates a paging interval at which it gets new data notifications from the base station. With multiple devices, you have to wake up N times as often because each of your N peers has a different paging time, and you’d have to also wake up even more for new peer discovery. Average one-way latency can’t be lower than half the length of your paging period when your receiver is off times the number of hops to the target.
Can you point me (a complete layman in this area) in a particular direction to get a bit more of a solid understanding of this area? I wouldn't even really know what search terms to use...
