5G and AI Self-Driving Cars


By Lance Eliot, the AI Trends Insider

Repeat after me, there’s 1G, and 2G, and 3G and more, plus 4G today and 5G is galore! I wax poetic about the future because 5G is on all our minds and we wait breathlessly for it to arrive. For those of you that are desirous of speed and have that incurable itch to get the fastest streaming you can find, I’m sure you are watching closely the upcoming emergence of 5G.

The Fifth Generation (5G) of wireless technology is about as big a deal as the invention of writing and the creation of fire. The telecommunications companies are incessantly teasing us with the prospects of what 5G can provide. You’ve likely seen ads that tout speeds at 100-times that of 4G, or maybe 30-times that of 4G, or some more sober writings that say it might be just 4-times 4G, or perhaps even slower than 4G. Say what? How can there be so many interpretations of what the speed difference is going to be?

Part of the problem is that the standard for 5G is still somewhat loosey goosey. Another problem is that just like for a car and MPG (Miles Per Gallon), as they say your mileage can vary, and so is the case for 5G. One setup for 5G can be materially different in performance from another. Generally, 5G architecturally makes use of the IEEE 802.11ac wireless networking standard and should, in theory, be a boost over the 4G standard of IEEE 802.11n.  Generally, it is claimed that 5G will be around 10-20 Gigabits per second (Gbps), while traditional 4G is perhaps around 100 Mbps, and 3G is around 384 Kbps. The proof though of actual 5G speeds will be in the pudding.

If we can really get the speed performance hoped for, you’ll see all sorts of exciting changes. The most obvious ones would be that you can do Virtual Reality (VR) and Augmented Reality (AR) in a manner that  is not especially feasible nowadays due to the slower wireless speeds of today. It is said that medical doctors will be able to perform surgeries remotely since the 5G speed will allow them to operate on a patient as though they were right there in the operating room with them. The Internet of Things (IoT) will shift from being a kind of novelty to becoming ubiquitous due to the 5G capability of making those portable wireless connected IoT devices more real-time capable in the receiving and sending of data.

5G is such a big deal that countries are fighting over the posturing of who will get there first. In the eyes of many, it’s primarily a China versus the United States kind of battle. Two titans, each wanting to get to 5G. Each expecting that it will transform their society. Each wanting not only the substance of advantages by 5G, but also the bragging rights too. It’s kind of like the Cold War era bragging over getting to outer space or perhaps even to the moon.

Some would say that China is already ahead of the United States. By leveraging its rather all-encompassing government control, China has seemingly gotten a massive alignment among its constituents, including the Chinese government, tech firms, businesses, and the like, all sharing the same 5G fever. Others would say that the United States, in spite of its more “freewheeling” approach to adopting 5G (some might say chaotic), will end-up getting there soonest as the all-mighty dollar is a sure means of creating momentum. These two horses on the race track are at times each perhaps slightly ahead or behind the other, but it seems a bit premature to be calling the race just yet.

Not all though is rosy about 5G. The city of Mill Valley, just outside of San Francisco and not far from the denizens of Silicon Valley, recently passed an ordinance to block or ban 5G from getting established there. Why would they stop the advent of such great wireless speed? Are they Luddites? They assert that there are potential health and safety concerns. You recall the ongoing debate about whether holding a cell phone to your ear can cause potential bad health consequences? Some believe that 5G might do something similar.

Part of why those concerned about 5G are troubled is that the path to 5G involves setting up lots and lots of small-cell 5G transmitters. The large towers that were put in place to get us to the 3G and 4G world are not what’s going to make 5G a reality per se. Instead, 5G will be mainly dependent on relatively small 5G transmitter devices that are going to be placed everywhere feasible. The good news is that their small size means that you can put them just about anywhere and you might not especially notice they are there. They’ll be on lampposts, on rooftops, on walls, on billboards, and so on.

There’s a word you ought to know to be techno-savvy about 5G, namely densify. If your normal vocabulary does not include the word “densify” that’s OK, since it is currently rarely used (soon, it will be a popular word!). It means to make something dense or compressed. The way that 5G works is that it generally can only go somewhat short distances and so you’ll need lots of them in a given area to have it work appropriately (they use higher frequency waves, while the 4G and 3G use lower frequency waves which can normally go further). Densification is coming. The telecommunications firms are going to want to put up the 5G communications devices on every inch and corner of where people are going to want 5G.

One of the biggest questions beyond the health and safety involves the cost factor. How much will 5G cost us? It’s going to take a lot of very costly technology infrastructure to get 5G put into place. The telecommunications firms are going to want to recoup those costs. There’s also the likelihood that the ongoing upkeep costs of 5G will be high, partially due the numerous 5G devices that will be laborious to keep functioning and repair/replace.

If the cost of 5G is “outlandish” then it will take some really pressing ROI (return on investment) to get someone to pay for it. It might be overly pricey for the everyday person to use 5G. Could be that only businesses with big bucks will be able to afford it. Of course, like any new technology, the odds are that as 5G becomes more pervasive, and with the competition gradually becoming more prevalent, we’ll see the pricing come down. Recall that when cell phones first appeared, it was a rarity to see one and only a select segment of society could afford it (remember, people used to wear it on their belt and it was considered a novelty and a conversation starter!). Now, we all seem to have cell phones and use them to our heart’s content.

For the moment, let’s put aside the cost factor and assume that we’ll have “affordable” 5G at some point (I’m not underplaying the cost aspects, just want to put it to the side for a moment to cover other facets, but please do keep in mind that cost will ultimately come to play in 5G, at least for sure when it starts emerging).

At a high-level strategic viewpoint, the main thing about 5G to be considering is that it will:

  •         Provide high bandwidth
  •         Involving faster speeds
  •         With more data at the same time
  •         Having a low latency (less delay)
  •         Using lower electrical power

If you have some kind of need for exchanging small sized packets of data, doing so very quickly, and without much delay, you ought to be considering 5G since that’s a sweet spot for it. 5G is supposed to provide sub one-millisecond latency, which is a nifty improvement in reducing latency.

Where might this kind of capability be needed?

I’d like to tell you about AI self-driving cars.

At the Cybernetic AI Self-Driving Car Institute, we are developing AI systems for self-driving cars. We also keep tabs on the latest high-tech trends and are determined to include any viable new-tech advances into the development of self-driving cars.

One such new high-tech is the emergence of 5G.

To consider how 5G might make a difference for AI self-driving cars, you need to consider what kinds of electronic communication needs might an AI self-driving car have.

5G Should Speed Over the Air (OTA) Updates

Let’s start with the OTA (Over The Air) updates for an AI self-driving car. Most of the auto makers and tech firms that are developing AI self-driving cars are making use of their own specialized cloud capabilities to be able to electronically communicate with their AI self-driving cars. Each such company is taking slightly different approaches, and so right now there’s not one super grand universal cloud for all AI self-driving cars (and, some doubt there ever will be, while others suggest that there should be, presumably as a means to have all AI self-driving cars learn from each other, but this also begs privacy and Big Brother issues).

For my article about OTA, see: https://aitrends.com/selfdrivingcars/air-ota-updating-ai-self-driving-cars/

For my article about privacy issues and AI self-driving cars, see: https://aitrends.com/selfdrivingcars/privacy-ai-self-driving-cars/

For my overall framework about AI self-driving cars, see: https://aitrends.com/selfdrivingcars/framework-ai-self-driving-driverless-cars-big-picture/

The OTA capability allows for both the uploading of data from the AI self-driving car and also the downloading of data and programs.

While the AI self-driving car is driving around, it is collecting all sorts of data from its myriad of sensors, including camera pictures and video, radar data, sonic data, LIDAR data, and so on. This can potentially be uploaded to the cloud. From this uploaded data, the auto maker or tech firm could use it for a multitude of purposes. One such purpose would be to use it for Machine Learning (ML) and try to improve the performance of their AI self-driving cars accordingly.

Another purpose would be to use it for analyzing driving patterns, such as where people go, how long it takes, etc. This could be used for purposes of improving our roadways and might be shared with certain governmental bodies. The data could also be used for business purposes such as being able to let retailers know how often you drive near to their establishment. There’s likely a treasure trove of data that could be either sold to third-party firms or analyzed for them. Money is likely to be found via the nature of the data collected. Whether the public will want this, or stand for it, we’ll need to wait and see.

For the debate about ML and AI self-driving car data, see my article: https://aitrends.com/machine-learning/machine-learning-data-self-driving-cars-shared-proprietary/

For idealism that some have about AI self-driving cars, see my article: https://aitrends.com/selfdrivingcars/idealism-and-ai-self-driving-cars/

The OTA also can download to the AI self-driving car. This is perhaps the most touted feature of the OTA. When your AI self-driving car needs an update to the software on-board, you won’t need to go to a local dealership to physically have an automobile mechanic load it. Instead, via OTA, the latest updates will be beamed into your AI self-driving car. This can be quite handy and would allow for your AI self-driving car to be updated with the latest Machine Learning results that the auto maker or tech firm has found, or add new features into the software, or fix bugs or errors that exist in the software.

For my article about possible software bugs in AI self-driving cars, see:  https://aitrends.com/ai-insider/software-neglect-will-impede-ai-self-driving-cars/

For the potential of AI self-driving cars to freeze-up, see my article: https://aitrends.com/selfdrivingcars/freezing-robot-problem-and-ai-self-driving-cars/

For the possibility of back-doors into AI self-driving cars, see my article: https://aitrends.com/selfdrivingcars/ai-deep-learning-backdoor-security-holes-self-driving-cars-detection-prevention/

Generally, the OTA is most likely to take place on a “batch update” oriented basis. In other words, right now, we’re all pretty much assuming that the OTA will happen when there’s a time and place available to do so. Some of the AI self-driving cars require that OTA only occur when the self-driving car is unmoving and parked. You would tend to think of this like an EV (electrical vehicle) which gets charged when it is unmoving and parked. Likewise, the OTA updates would normally occur when you AI self-driving car is perhaps parked at home in your garage, or maybe when it is parked at work.

Part of the reason for doing the OTA on a batch-oriented basis is due to the electronic communication needed with the AI self-driving car. Trying to electronically communicate with an AI self-driving car that’s going 75 miles per hour on the freeway is likely problematic. The connection is bound to be spotty, slow, and get disrupted. Instead, waiting until the self-driving car is in a nice quiet place and not underway would seem handier.

There are other factors involved too, such as even if the OTA could happen while the AI self-driving car is in motion, it would seem untoward to consider making updates to the on-board AI system during it’s crucial efforts of driving the car. You might be willing to download updates to the AI self-driving car, but you’d be safest to wait to actually have those updates installed, doing so when the AI is otherwise unoccupied with keeping your car on the road and you safe from getting killed.

The 5G can help out with the OTA.

The amount of time required to upload or download data from the cloud to the AI self-driving car could be lengthy under a 4G or 3G scenario. The hope is that 5G will be much faster and therefore the OTA time will be lessened. We don’t know how long it’s going to take to normally do OTA on an ongoing basis for most AI self-driving cars, but at slower wireless speeds of today and possibly with large updates being downloaded or the voluminous data being uploaded, it could presumably at times take hours to do (5G faster speeds might reduce it to minutes).

So, score one point for 5G on helping out with OTA for AI self-driving cars.

There’s another way in which 5G might help.

We pretty much assume that people will start riding in their cars a lot when they have an AI self-driving car. It’s going to be handy to have an electronic chauffer that will drive you wherever you want to go. When you drive to work in the morning, your hour-long driving commute will no longer involve you driving the car and so you’ll have that hour to do something else instead. I suppose you could take a nap, but maybe instead you’ll want to be entertained, such as streaming the latest baseball game underway or watching a movie that’s just been released. Or, maybe you’ll use the time to take an online college course.

Also, AI self-driving cars are going to boost ridesharing. People will turn their AI self-driving car into a money-making machine. While they are at work, they’ll be renting out their AI self-driving car. While people are sleeping, they will rent out their AI self-driving car. We’re heading toward a mobility-as-a-service economy. When consumers opt to take your AI self-driving car for a ride, what will they do while inside your AI self-driving car? One answer is that they might like to be entertained by having streaming video or audio.

For the non-stop aspects of AI self-driving cars, see my article: https://aitrends.com/selfdrivingcars/non-stop-ai-self-driving-cars-truths-and-consequences/

For the burgeoning of ridesharing services via AI self-driving cars, see my article: https://aitrends.com/selfdrivingcars/ridesharing-services-and-ai-self-driving-cars-notably-uber-in-or-uber-out/

All of that in-car “live” entertaining requires electronic communication. Of course, you could have pre-stored movies or other pre-taped material instead, but I’d bet that humans will want fresh and live content while in an AI self-driving car. With the advent of 5G, it will make that streaming even more glorious. At the speeds of 3G and 4G, the streaming might be rather stilted and broken. The hope is that 5G would make it seem smooth and watchable.

We now though need to consider the reality of how that 5G is going to happen.

If your AI self-driving car is parked at the office or at home in your garage, the use of 5G is going to be quite likely and thus the OTA will be fine. It would make sense that you’ll likely have 5G transmitters in your home, or at the office. But, the live streaming of entertainment and other such video and audio, while the AI self-driving car is in motion, that’s going to be tricky to arrange. Recall that I mentioned the 5G goes usually only short distances. How will that happen while you are cruising on the freeway?

The odds are that we’ll need to have lots and lots of 5G devices placed along our major highways and byways. Thus, as the AI self-driving car is rocketing down that freeway, it is moving right near and past a 5G transmitter and another one and another one. This will require some clever handling of hand-offs from device to device and still keep the moving car connected in any fluid fashion. This tough problem does though seem technologically feasible and can ultimately be worked out.

Besides in-car entertainment, AI self-driving cars are likely to have a zillion Internet of Things (IoT) devices either included into the self-driving car by the auto maker, or by having been added-in, or by human occupants that carry into the self-driving car their IoT devices such as their fitness wristband, their IoT connected jewelry, their smartphone, their tablet, etc.  The 5G capability, once again via lots of transmitters along the roadside, would allow those IoT devices to ping and play.

Here’s another way in which 5G can help, and it actually deals with the driving of the self-driving car.

AI self-driving cars are likely to have V2V (vehicle-to-vehicle) communication capabilities. This will allow one AI self-driving car to electronically chat with another AI self-driving car. Suppose there’s an AI self-driving car a mile up ahead of your AI self-driving car, and it has spotted a car accident. It could communicate to your AI self-driving car and forewarn it about the accident. Your AI self-driving car might then start to slow down or maybe take an alternative route that goes around the accident scene.

This V2V capability is still being worked out. There are evolving protocols about what is communicated and how it is communicated. We need to deal with the dangers of false V2V that tries to trick an AI self-driving car. There’s also the concern that a DoS (denial of service) type of attack could happen by someone swamping your AI self-driving car with tons of meaningless messages via V2V. But, anyway, the other factor is how will the V2V electronic communication actually take place.

One approach is that the V2V could be done not just directly from one vehicle to another, but maybe it first goes to a 5G device near the roadside, which then relays it to the destination car that’s nearby. Thus, the AI self-driving cars could be bouncing their electronic messages off of the 5G devices near the roadside and then other AI self-driving cars could be picking up those messages.

This also brings up the rise of V2I (vehicle-to-infrastructure). We’ll gradually be having our roadway infrastructure outfitted to be “smart” in the sense that a bridge might be able to electronically transmit that it is overly crowded and to be avoided. Or, the traffic signal at the intersection near your home might send out an electronic signal, which would be handy for an AI self-driving car, rather than solely being dependent on displaying a green-yellow-red light. AI self-driving cars will be seeking out the V2I data so as to properly route the self-driving car and be able to avoid traffic jams or more align with traffic conditions.

Once again, 5G could help out.

When I say that 5G might be able to help with the V2V and V2I, it is especially notable that we’re talking about 5G because of the low latency involved. It’s one thing to transmit an electronic message, but if there’s a delay in it getting received, the delay time with an AI self-driving car could be crucial. A split second might make the difference between the AI making a determination to brake the self-driving car and avoid a terrible accident. If the transmission is coming from a 3G or 4G transmitter versus a 5G, the latency could be the difference between life and death.

For the IoT aspects of AI self-driving cars, see my article: https://aitrends.com/ai-insider/internet-of-things-iot-and-ai-self-driving-cars/

For my article about the crucial timing aspects of AI self-driving cars, see: https://aitrends.com/selfdrivingcars/cognitive-timing-for-ai-self-driving-cars/

We might have more than just a slew of 5G transmitters scattered throughout our roadways. We might also have some serious computing power that’s also nearby the roadways. This is referred to as “edge computing” and it’s a somewhat new trend toward putting computer processing closer to wherever the action is.

In the case of AI self-driving cars, the notion for edge computing is that an AI self-driving car might be able to do more of a real-time interaction with the auto maker or tech firm cloud if it had the processing and data sitting closer to where the AI self-driving car is (in this case, zipping along on the highway).

Nokia Conducting Trials with 5G and Self-Driving Cars

You might want to keep your eye on the work being done by VTT Technical Research Centre of Finland and being done with Nokia, in which they are doing tests and trials of using 5G with self-driving cars and edge computing. Using the VTT self-driving car called Martti, they are doing some interesting explorations on a vehicle test track in Sodankyla. The 5G though is still in prototype stages and thus we’ll need to wait and see what happens once they’ve got some “real” 5G to tryout. They are also experimenting with being able to dynamically select whatever radio signal is actually available, meaning that if 5G is available it uses 5G, but if only 4G is available it contends with 4G, and so on. The overall project, known as 5G-Safe, involves considering autonomous driving along with time-critical aspects while the car is on the move.

When you think about an AI self-driving car, it’s helpful to be aware of the levels of self-driving cars. A true self-driving car is considered Level 5, it’s the topmost ranking of a self-driving car and refers to the aspect that there is no need for a human driver. Indeed, there’s usually no brake pedal and no gas pedal, and no steering wheel. Self-driving cars at less than a Level 5 are expected and must still have a human driver present and so also have the pedals and steering wheel in the car. In the less than Level 5 self-driving cars, there is a kind of co-sharing of the driving task, which has some significant disadvantages and provides inherent dangers.

For more about the disadvantages of co-sharing the driving task, see: https://aitrends.com/selfdrivingcars/human-back-up-drivers-for-ai-self-driving-cars/

For the levels of AI self-driving cars, see: https://aitrends.com/selfdrivingcars/richter-scale-levels-self-driving-cars/

For my article about edge computing, see: https://aitrends.com/selfdrivingcars/edge-computing-ai-self-driving-cars/

In theory, an AI self-driving car of a Level 5 is considered fully autonomous. It requires no human driver.

Please be aware that some are stretching this a bit by providing a remote “pilot” human driver in case the AI goes awry or gets stuck, but I’d dare say that most purists would contend this is not a true Level 5 per se and not fully autonomous, not at least in the sense that it is able to drive the car without any human intervention needed.

Some assert that using a remote human “pilot” is a bad idea overall, since it really is just a crutch and won’t motivate us as AI developers into fully developing the AI be a truly fluent and self-sufficient driver of a car.

For remote piloting of AI self-driving cars, see my article: https://aitrends.com/selfdrivingcars/remote-piloting-is-a-self-driving-car-crutch/

Let’s though put aside for the moment the notion of a remote human “pilot” and pretend it is out of scope for this discussion.

This brings us to one of the more interesting and challenging questions about an AI self-driving car, namely whether it should be entirely autonomous with respect to itself, or whether it is OK to consider having it interact with other AI that might not be on-board of the self-driving car (AI that is “off-board”), doing so in the act of actually driving the car.

Allow me to elaborate.

Suppose your AI self-driving car is going along on the freeway. It’s doing its thing and driving the car just fine. It’s a Level 5 self-driving car. There are no humans involved in the driving task. All of a sudden, right there in the middle of the freeway a family of ducks is waddling along (this really happened here in Southern California recently – you’ll be relieved to know that the ducks survived, and no humans were hurt in the process of avoiding the ducklings).

Anyway, the AI of this particular self-driving car has perhaps never before encountered ducks on the freeway. What should it do? It might have some general principles to fall back onto such as don’t hit objects on the freeway. It might try to compare the images of the ducks to other images of animals, such as dogs and cats, which perhaps it’s Machine Learning had already figured out what dogs and cats look like and act like. But, the ML ascertains merely that these ducks don’t seem to be dogs or cats.

The AI might have then some generalized plan that if there is a series of moving objects, and they aren’t identifiable per se, it should try to avoid them. This though might be tempered by the aspect that if the objects are small, and if avoiding them might be a risky maneuver, it is permitted to run over the objects, which is perhaps the safest choice in some cases. For example, if a tumbleweed is blowing across the freeway, it might be safer to strike the tumbleweed than it is to suddenly come to a halt or make a dangerous swerve to avoid it.

Under the logic that these are small objects, and that they are comparable to tumbleweeds, the AI might determine that the “right” course of action involves running over the ducks. Horrid! Can you imagine the headlines: “AI Quackery Smacks and Kills Ducks on Freeway” (or something like that).

Here’s a possible alternative. The AI of the self-driving car, being somewhat stumped about these small objects, opts to seek out a second opinion. It opens an electronic connection with a nearby edge computing device, doing so via 5G, and shares with the computing device what it knows about the objects. This computing device, placed there by the auto maker or tech firm that made the AI system of the self-driving car, quickly looks up and discovers that ducks have been seen before and thus it knows what to do about ducks on the freeway. It shares this piece of “knowledge” with the on-board AI system and the AI self-driving opts to avoid hitting the ducks. The ducks are saved!

We have then three kinds of autonomous AI self-driving cars:

  •         Standalone autonomous
  •         Exo-dependent autonomous
  •         Exo-augmented autonomous

The standalone autonomous is an AI self-driving car that is not setup to do any kind of AI-to-AI related co-sharing of the driving task. It is entirely standalone. No phone an AI-friend possibility.

The exo-dependent autonomous is an AI self-driving car that uses an external AI-capability to help undertake the driving task, and for which the AI self-driving car is dependent upon the external AI to be available. In other words, the external AI is vital to the driving task and the AI of the self-driving car is not standalone autonomous. There are driving situations for which the on-board AI must have access to the external AI in order to perform the driving task.

The exo-augmented autonomous is an AI self-driving car that uses an external AI-capability to help undertake the driving task, but the on-board AI doesn’t necessarily need the outside help. It will use it if the help is available, and give it due consideration, but otherwise even without the outside AI assistance the on-board AI is able to undertake the driving task.

Keep in mind that the exo-dependent and exo-augmented variants seem like the remote “piloting” that I mentioned earlier, but this is not quite the case because the conventional notion of remote “piloting” is that a human is the remote pilot. In the truly autonomous AI self-driving car, we’re still keeping the human out of the picture, and instead simply saying that we might distribute out the AI.

In the example of the ducks on the freeway, the AI self-driving car opted to use the localized edge computing that had local-AI pertinent to this particular AI system. We’ll assume in this situation that the on-board AI of the self-driving car was of the exo-augmented autonomous style, and could have proceeded without having accessed the off-board AI. Sadly, if the on-board AI had not opted to consult the off-board AI, and in this example, it might have cost the ducks their lives. But, I don’t want you to assume that just because there’s an off-board AI that it means the off-board AI will always somehow save the day.

You could argue that the off-board AI might make things worse, at times. It could be that the off-board AI makes a determination that if it had not been consulted then the on-board AI would have made the “right” choice. It could also be that the use of the off-board AI chewed up time, during which the on-board AI, if it had made a choice sooner, would have been better off, rather than having waited until the off-board AI was consulted.

For more about federated Machine Learning, see my article: https://aitrends.com/selfdrivingcars/federated-machine-learning-for-ai-self-driving-cars/

Trying to distribute out the AI of the AI self-driving car is only going to be viable if you have very high speeds and very low latency. Any significant delay of the on-board AI to the off-board AI, and back from the off-board AI to the on-board AI, kind of undermines the whole deal. You might be able to use the off-board AI for broader planning purposes, but any kind of in-the-moment needs would not be viable unless you have the appropriate electronic communication capability, which might well be 5G.

I’d like to also point out that we can extend the distributed AI aspects to include the other nearby AI self-driving cars. I had mentioned before the V2V aspects of AI self-driving cars. Presumably, the on-board AI of the self-driving car could have potentially asked another AI self-driving car a question about the object on the freeway. It is quite possible that another brand of AI self-driving car might have already known about ducks, and so it could have told the upcoming AI self-driving car what they are and offered a suggestion of what to do.

5G, it is a galore. Exciting times are a coming. For AI self-driving cars, it bodes well that 5G is emerging. With it, the nature of the OTA speed and timing might be dramatically improved. The in-car entertainment and use of IoT might be vastly improved. The federated or distributed AI might be feasible, becoming so because of 5G. It’s handy timing that just as AI self-driving cars are getting to a stage of reasonable feasibility, so too will 5G be coming along. They are two brothers, each aiding the other (5G obviously can enhance AI self-driving cars, and a justification for 5G would be that AI self-driving cars can be enabled or improved via the advent of 5G).

Let’s be clear though that 5G is still being developed and implemented. We don’t know how good it will be. We can’t say how long it will take to get perfected. And, it will certainly be costly to put it in place, especially if you are anticipating having squillions of them lined along our freeways and highways.

Let’s not hold our breath for 5G. All the work on AI self-driving cars should be proceeding for now under the assumption there is no 5G per se. That being said, we should be working towards the next iteration of AI self-driving cars, wherein 5G is available. And, dare I say, we can even start on designs of AI self-driving cars of the further future that will use 6G. I know it’s a bit farfetched to already have my eye on 6G, but hey, it’s never to early to start planning for the future. Soon enough, 5G will become tiresome and worn out, and we’ll be all signing the praises of 6G.

Copyright 2018 Dr. Lance Eliot

This content is originally posted on AI Trends.