6G: Connect Us!
On February 21, 2019, @realDonaldTrump tweeted, “I want 5G, and even 6G, technology in the United States as soon as possible. It is far more powerful, faster, and smarter than the current standard.” Thus the POTUS released his high-level 6G design goals.
It is so early in 6G’s lifecycle: the goals are unclear, there is no broadly agreed specification, and many enabling technologies are still in research labs. Since 2018 significant government and public/private 6G research partnerships have been launched in the US, Finland, South Korea, China, and Japan. They are studying the optimal spectra, electronics, user needs, and business models for that next generation.
A decent first assumption is that 6G will be in the continuum of evolving cellular network standards that see 10X to 40X improvements in bandwidth every ten years.
While 5G, millimeter waves are achievable with technology that we have and understand, 6G’s terahertz frequencies (THz) involves solutions for creating and receiving the signals that is truly experimental. One terahertz, or 10^12 times per second, is an impossibly large number. If you stack 10^12 golf clubs end to end you will reach Saturn. THz signals are sometimes referred to not as waves but as terahertz radiation. It’s technically close to the infrared radiation spectrum: think of heat, and night vision. Designing electronics to create and receive the signals will be an adventure.
Assuming 6G’s performance follows this path, what could you do with it? Is 6G just specmanship like new software that coaxes your Tesla from 0 to 60mph in 2.5 seconds instead of 2.8, or will it solve real problems? How about telepresence.
Our five powerful senses conduct all of our interactions with the physical world. It is astounding how much interpersonal collaboration and productivity we have achieved in the internet age with just audio and flat, 2D video, engaging a thin slice of what our senses perceive. Even the best virtual reality (VR) system leaves touch, smell, and taste to be completed by our imaginations. Video calls from the road with our families are better than voice alone, but still leave emptiness.
What type of connectivity might it take to give us a rudimentary telepresence experience, one with a 3D holographic image and the ability to touch a remote environment?
Basic Telepresence: Holograms and Touch
Holography was developed in the late 1940’s by Dennis Gabor, who later won the Nobel Prize in Physics for his work. Renditions of holograms have appeared in numerous movies, perhaps most memorably in the Star Wars saga, and similar 3D imagery has captured headlines through events like the performance of Tupac Shakur’s “ghost” at Coachella in 2012 (which was not a hologram but a 3D animation projected onto an angled piece of glass). The psychology is clear: the more realistic, the more compelling.
Real holograms have been the subject of much research and demonstration in the past ten years, particularly with regard to the bandwidth required to transmit them. In 2011, researchers at the Data Science Institute in Singapore determined that bandwidth of roughly 10Gb/sec was required to communicate an HD hologram of a 3-inch-tall full color object at a modest frame rate of 60 frames per second (fps. Frame rate is how fast your computer screen is refreshed, and 60fps is so slow that most humans can see the screen flickering at that rate. 72fps is a reasonable minimum target).
Huawei estimates that 1Tb/s is needed to transmit useful holograms. Nokia and Orange demonstrated hologram transmission over a 1Gb/s 5G connection in 2018, but the viewers had to wear Oculus VR headsets. They highlight how the amount and nature of the edge computing used to recreate the image can impact the bandwidth needed to send it. Given that Nokia and Orange have a strong vested interest in promoting 5G adoption, that their demo was under controlled conditions, and that other analyses point to substantially greater bandwidth being needed, I conclude that 5G will be insufficient for usable telepresence holograms. The performance requirements are comfortably in the range of a supposed 6G network.
What about the sense of touch? When you buy new clothes, especially with a new fabric or weave, don’t you want to feel its weight and texture? What’s required to transmit tactile information? My own analysis, shared in a note at the bottom, is that 5G will likely be capable of carrying full tactile information for your two hands. The rapid growth of machine learning, specifically pattern matching capability, adds a tantalizing twist to any telepresence analysis. Suppose a pattern-matching AI is capable of extracting complete tactile information from a 2D video so that no additional data is required? It’s a ripe area for exploration and innovation.
We will likely have solid answers for the transmission of a holographic, tactile 3D world in two years. The ten-million-dollar ANA Avatar X-Prize, a competition ending in March, 2022, now has 77 global teams competing to realize the design of a visual and tactile telepresence system. Why would an international airline, ANA, sponsor a high-stakes telepresence competition? I had a conversation with one of ANA’s senior marketing staff shortly after the competition’s announcement at SXSW 2018 and her explanation reflects a company that deeply understands their value to their customers. She noted they are in the business of connecting people with other people and places. ANA sees telepresence not as a threat but as an opportunity to do more for their customers.
Such a level of connectivity will exacerbate questions about work visas and immigration. If your team in Buenos Aires can holographically appear in meetings in San Francisco with compelling presence, rolling about the office in telepresence platforms, will they need work visas? Will ICE agents bust in and pull their batteries much like police cut power to Chief Keefe’s holographic concert appearance in 2015?
What 6G Might or Might not Do: It’s a Matter of Time
Network latency receives much less popular attention than bandwidth but it may be the greatest opportunity for compelling progress beyond 5G. Time, distance, and the speed of light will have everything to do with squeezing more benefits out of the airwaves.
In our age of machine-to-machine communications time is literally money. In virtually all financial markets the saying is, “Milliseconds mean millions,” and as soon as the humans and machines shifting trillions of dollars around the globe can access something with less delay than 5G, they will. While the best-case latency for 5G is stated to be 1ms a typically expected delay is 10X that, 10ms: that’s lots of $millions.
Latency is also pivotal in how you and machines communicate and interact with remote environments. Suppose you want to virtually see an environment, to turn your head and have a remote camera eye around for you. A common VR guideline is that the lag between when you look in a new direction and the scene changes is about 13.9ms, equivalent to 72 frames per second of video. Slower than that leads to nasty side effects including dizziness, nausea, and headaches. Because the command has to go out and the image has to come back, 7ms is the latency target. 5G will be hard pressed to reliably deliver it; there is opportunity for the next standard.
Controlling an industrial machine capable of moving at beyond-human speeds is even more exacting. 3GPP, a standards group for wireless telephony, trims the latency target to 0.1ms for radio control of industrial robots. That is far beyond even the best-best case scenario for 5G and is a compelling use case for an aggressive 6G specification.
Physics That 6G Can Not Overcome
It’s intriguing that consideration of 6G keeps coming back to physics. Electronic signals travel through the air at the speed of light. They are in fact light, just at a much lower frequency. You can’t get a signal from one point to another any faster, until of course we fully control whatever we think quantum entanglement is (I’m not holding my breath).
You will read a lot about remote or “cloud control” of autonomous cars. An autonomous car is just a robot, so assume the 3GPP spec. of 0.1ms total there-and-back latency is needed for safe control. Light will travel a little under 30,000 meters, or 18.6 miles in that amount of time, a maximum separation of machine and controller of 9.3 miles (unrealistically) assuming instant calculations and zero additional delay anywhere in the network. Among the closest satellites to Earth are the 60 Starlink units that SpaceX launched last November in a low orbit of 217 miles (350Km). They’re not anywhere near close enough.
Critical reactions for autonomous cars may never be remotely controlled unless we invest in 100% available data processing centers within 9 miles of every point on every road everywhere. I’m not holding my breath on that, either, but as with telepresence the opportunity for creating more powerful local processing is abundantly clear.
In real use your wireless signals will not travel a straight line: they will go from you to a tower then a satellite dish then a satellite followed by another dish and another tower and finally to your robot/colleague/video-server, then back. Whether it’s telepresence, telesurgery, or telefinance, when time is of the essence the speed of light will tell you the fastest that it can possibly happen.
Time Isn’t The Only Thing
Some vital human needs do not require anything close to the speed and latency of even 5G. Performance requirements for communicating blood pressure, pulse, blood glucose level, acceleration (falling), and even EKG graphs are modest: what they need is superb reliability, availability, and security. A dropped call is annoying; a dropped vital monitor may be life-threatening.
Reasons for unstable connections include weak signals due to poor coverage, handoff errors between towers, and towers being overloaded with too many connections. A major benefit of 5G is the increased number of connections available per cell: 6G network development has an opportunity to do even more. A key purpose of portable health monitors is mobility. For those dependent on them ultra-reliable coverage means freedom.
Security is of course both a network and an application issue. Considering only the network, 6G’s potential use of THz frequencies and as yet undeveloped electronics will make physical signal interception the domain of only the most highly sophisticated and well-funded technologists. Signal security will be an increasingly important issue as we explore a new way of interacting with our environment. Consider that connecting machines with human minds, the so-called Brain-Computer Interface (BCI), is poised to make major advances in the next ten years.
When Will You Invest in 5G?
I will not be buying a 5G-capable mobile phone this year. With an extra cost on the order of several hundred dollars and virtually all of my mobile phone use being indoors, the investment case is pretty weak.
In 10 years, Joe Buck may introduce Super Bowl LXIV as “the first Super Bowl featuring 6G ultrawideband!” I’d love to be watching from my couch exchanging high-fives with family and close friends in distant places. Over the long-haul one can’t predict exactly how 5G will be used, what 6G will be, or their ramifications: aren’t most important technologies like that?
It is safe to predict that with substantially faster and denser connectivity we will produce more while using fewer resources, and interact with each other more conveniently and authentically.
What will that enable you to do?
Note: Calculations on communicating tactile information
- This 2018 paper in Micromachines notes that texture receptors sense at rates between 300Hz and 1KHz. Additionally, summing all forms of tactile sensors in the human fingertip, their density is 250/cm^2. Assume that extends across the entire surface of your hands and estimate that area to be 22cm x 18cm so ~400cm^2: that’s ~100,000 receptors.
- Assume 16 bits of resolution needed per receptor thus 1.6Mb of binary information needed, at 1Khz, so ~1.6Gb/s of raw tactile information. With good data compression that might be reduced by a factor of 10 to 100 thus reasonably within the capability of 5G.
- I’m ignoring any analysis of required hardware because such fine tactile sensors and systems are largely in the research stage. Hardware is hard, but when it’s available and there’s strong demand the cost will plummet.