A First Use Case of 5G
What is one of the first relevant and currently realistic use of 5G? Outdoor cameras are probably underestimated favorites.
The fifth Generation of wireless communication, which we know as 5G, largely outperforms 4G in both latency, the transportation time of the information, and capacity or bandwidth, i.e., the maximum amount of data that can be transported. Comparatively, when the 4G latency is 40 milli-seconds (ms), the latency of 5G is 1ms. 5G supports 1000 times more users per unit area than 4G and the download time in 5G varies around 500 megabytes per second (Mbps) while 4G download time is around 15 Mbps.
The capacity and latency of 5G networks enable the deployment of wireless technologies previously unpractical. The high capacity allows for more cellphones and Internet of Things (IoT) objects on the network. We’ll transition from smartphones to smart cities. The extremely low latency of 5G is paramount in critical applications. For instance, this will help prevent collisions with drone delivery or autonomous cars, the forthcoming Internet of Vehicles (IoV). This will also enable outdoor digital applications, like augmented reality and virtual reality. There will be glasses that seamlessly add digital layers of knowledge to our perception of the world. We’ll watch films and gaming in 3D anywhere at the speed of light.
Yet, this is very speculative. Is the timing right? Are these technologies going to be in the streets in 2022 or in 2030? How long will it take for autonomous cars to work? To be approved? To take over the car market? This looks like science-fiction movies, as in the Ready Player One world. We may believe that the future will look like that. Yet, how will it practically unfold? What are the small steps that will lead us to that future? What will be the first valuable applications funding the deployment of that technology? When looking back at our past two decades, the dot-com bubble enthusiasm was a justified oracle of the future. People felt that the internet would start a game-changing new era, the information era. Changes have been dramatically fast; yet, they remain more gradual than investors projected hopes’. This is why it is interesting to careful stake out the potential first usages of 5G.
China is the current leader in 5G and has installed 85% of 5G base stations worldwide. Let us learn from them. For that, the China Academy of Information and Communications Technology (CAICT) offers solid points of view. In this report, they detail a variety of uses cases of 5G, mainly in an industrial environment. 5G is beneficial for fast drone inspection of industrial plants, unmanned operations in coal mining, containers’ optimization in harbors, IoV, Healthcare, etc. Somehow it is not surprising: find situations where you cannot have optic fiber cables and where low latency and high capacity are necessary.
Yet, there is a usage of 5G that is not mentioned: surveillance cameras on the network. It is an immediate and relatively straightforward application of 5G technology. Why not portray the usage a bit more? Well, as opposed to other applications, it is arguably perceived negatively (just look at the sentiment analysis of “surveillance cameras” or even “outdoor cameras”). Yet, outdoor cameras are predicted to be the most significant objects in the 5G IoT in 2022 and 2023. To get a confirmation of the trend, please take a look at the google trend result for the research “outdoor cameras”.
Current cameras in major cities are currently either connected to closed-circuit systems (CCTV) or to wifi network like IP cameras.
With the 5G large bandwidth, one simply needs to power a camera and send its video stream back to a central server. Besides, the low latency and large bandwidth enable sending much higher quality videos, which notably improves the stream processing algorithms’ performance. For instance, face recognition could be less prone to errors; see this report at the section “5G+smart security: efficient city protection anywhere, at anytime”.
There can be as many cameras as there is street lighting. The movie “The Circle” would not be science fiction anymore.
But, what would we do with all the cameras? How can we sort the information? We would need many humans behind the camera to exploit that information? Plus, it is not soon that European or North American countries will allow face recognition, so how beneficial all these cameras can be?
In computer science, a recent and significant breakthrough is that some algorithms are capable of automatically recognizing action in a video stream. These A.I. algorithms can detect, in real-time, what a person is doing on a video, from throwing the garbage to observing a dog illegally peeing on the street. These algorithms need a lot of computing power to run efficiently. The computation cannot be done locally on the camera; the video stream needs to be sent to a server. 5G enables such systems. These algorithms are the equivalent of search engines in the early days of the internet. Without them, we would have drowned under the quantity of information to sort. Similarly, with action recognition algorithms, surveillance centers do not drown under an overwhelming number of camera screens. Five years ago, 5G would have had very little use for camera surveillance networks.
And the system needs not to be fully automatic nor perfect. Algorithms indicate to a human operator where to look (with some potential mistakes). Algorithms act as external brain attention mechanisms that enable a few human operators to have an eye everywhere and anytime in a city. And, ultimately, this would dramatically increase the coordination with police enforcement. Here, we apply no judgment; it is simply a logical thought that leads us to believe that surveillance cameras will initially be one of the most significant profiteers of 5G.
But isn’t science-fiction too? Would people agree to such systems? In Paris, there is already an automatic traffic fine system with cameras. You need not be arrested by a policeman or caught by a speed limit camera anymore. My mother got two automatic fines, with no proof attached, for impinging on the bus traffic lines. Similar products to cameras in the 5G network are already in the market!
But my mother was identified via her car’s license plate. Face recognition algorithms are to humans what license plates are to cars. Face recognition will probably be part of Europe and North America’s future, but when? I wouldn’t believe it will be there in 2022 nor 2023. So, does the whole argument fails? Aren’t surveillance cameras on the 5G networks augmented with action recognition algorithms useless without face recognition or some sort of identification?
Action recognition is just one example of the huge automation possibility recently achieved by A.I. algorithms on computer vision tasks. It is not very hard to automatically track an individual from one camera to another provided sufficient coverage. It becomes easy to coordinate with police enforcement to arrest criminals by following them from camera to camera. In social turmoil contexts’, such as the yellow vests protests’ in Paris, it seems reasonable that governments will seek to embrace such surveillance systems.
Overall, although it is not much publicized, surveillance cameras in the 5G networks seem to gather all the elements that make it a feasible market opportunity supporting the development of 5G networks. It does not change much the exciting city landscape; it does not require new technology, and nearly all big cities in the world are potential clients. What do you think?
