In practice, numerous AI technologies have been applied to wireless communication, from electronic engineering, such as IC design, electromagnetic simulation and computer-assisted optimization, to resource management such as the dynamic management of base stations and the management of spectral and storage resources, and to information processing such as security and authentication, detection and prediction. In some ways, AI has become an integral part of wireless communication and will be even more important for 6G. The final form of AI-defined wireless communication will make the earth an intelligent planet. It relies on realizing an integrated AI system from the device to the entire network.
Supplements to AI
Edge computing is a distributed computing solution. Researchers believe the data transmission resources required and latencies introduced by the data transmission between terminals and the central servers will be unacceptable. In the following decades, Industry 4.0 is expected to evolve toward a massively distributed production organization, with connected products (i.e., with communication ability), collaborative AI robots, integrated manufacturing and logistics management.7 By distributing data producing devices between the cloud and the edges of the networks, edge computing enables distributed cloud applications while providing higher speed and reducing latencies. It could support more personalization and the management of personal data as well.
6G must be intelligent and world-connected, including one or several cores and countless distributed edge devices. Utilizing big data, deep neural networks and edge computing in conjunction with game theory and self-evolution will provide the “brain,” which along with the 6G wireless communication network will ensure the internet of everything, anywhere and anytime.
“Cloudified” and “Foggified” New World
Cloud computing refers to the wide area network or local area network including hardware, software, network and other unified resources to achieve data computing, storage, processing and sharing of a managed technology.
The concept of fog computing was proposed in 2011. It is a combination of weak and more decentralized computers, which infiltrate all smart devices, including mobile devices, personal computers, smart cars, and various intelligent objects in people's lives. No matter how weak the capacity of a single node, it plays a role in the entire network. Fog computing is an emergent architecture for computing, storage, control and networking that distributes these services closer to end users along the cloud-to-things continuum.
Compared with cloud computing, the architecture of fog computing is more distributed and closer to the edge of the network. Fog computing decentralizes data, data processing and applications on devices to the edge of the network, rather than storing them entirely in the cloud. Data storage and processing rely more on local devices than servers. Combining cloud and fog computing will create a new generation in line with the "decentralized" goals of 6G.
AI-Defined Wireless “Blockchain” Networks
Distributing data producing devices between the cloud and network edges provides higher speeds and reduced latencies, while Blockchain, a decentralized distributed database, allows non-trusting members to have distributed peer-to-peer connections verifiably without a trusted intermediary.
For wireless communication, an AI-defined wireless “blockchain” network an alternative dynamic sharing technology with reduced cost and without the need for a centralized database to support access and sharing. It enables the nearby wireless devices to build a quick-connect dynamic temporary network, to achieve data sharing, computing power sharing, network enhancements, sensors or other device sharing.
Materials, Electronics and Physical Science
Novel material technologies such as graphene open the door to breakthroughs in room-temperature superconductors and battery technology. Heterogeneous integration makes it possible to fabricate microwave analog and digital circuits on one chip. New material science is the key; it provides the possibility to realize technologies that were once only theoretically possible.
Moore’s law states the number of transistors in a densely integrated IC doubles about every 18 months. Moore's prediction proved accurate for several decades and has been used in the semiconductor industry to guide long-term planning and to set targets for research and development. However, IC processing is predicted to reach a limit of 3-5 nanometer (nm) by 2025, after which Moore's law may be no longer hold. It is generally believed that when the device size is close to 5 nm, charge carriers are subject to quantum mechanical effects, and current classical theories break down.
6G coverage will not be limited to the earth’s surface, but will extend into deep water, deep underground, near-earth space, beyond our solar system and into deep space. Quantum information science is the most promising way to realize space level real-time communication. Quantum science is also the key to extending the life of Moore’s law.
Ethics
When developing science and technology, it is important to consider philosophies and ethics as early as possible. These are just a few examples:
The Trolley Problem
Philippa Foot proposed a famous moral psychology dilemma. He described a spectacle of an uncontrolled trolley in which the driver must choose whether to hit one person on the rail or five innocent bystanders. It is simple if the decision is made by an individual; no matter what the choice, it can be explained by a choice made in the heat of the moment. However, when it comes to AI, it becomes complicated since the code is previously programmed. That means the AI's behavior is completely predictable, and therefore the risks must be evaluated.
Gambling and Free Self-Evolution
Game theory and free self-evolution of AI leaves open the possibility for unpredictable behavior. Self-evolution of AI is considered both promising and dangerous. All intelligent things are ultimately realized through computing, and the fundamental theoretical model of the computer is a Turing machine. The "gene" of the Turing machine, and hence AI, is Turing code. Since Turing code is the "gene" of AI, it is the key to realizing AI self-evolution. When AI evolves enough to alter Turing codes and even create new algorithms, it might evolve in its own way, with the possibility of growing out of control in its self-evolution “game.”
Data Discrimination
With strong AI and big data analysis, people live in a panoramic prison under 24 hour monitoring. AI and big data, including age, income, location, marital status, employment, consumption habits and hobbies of each individual; as well as information disclosed in histories of purchases and browsing, security personal information collected from email, phone calls and social media provide differentiated individual information. Sometimes, products or services pushed by advertisers based on big data mining and AI are so suitable as to stimulate the desire to buy. So, to a certain extent, AI and big data know you better than you know yourself. The problem is that while AI and big data know you, they don't necessarily love you. Instead, they use what they know about you to achieve their own purposes, for example, by increasing an airplane ticket price based on travel habits.
Replacement Crisis
Beyond the replacement of handicraft by the machinery manufacturing industry, AI and its derivative technologies have the potential to make people in all fields unemployable, including manual workers, mental workers and even politicians. It is important to make use of AI to enhance the abilities of human beings, rather than the simply replacing their livelihoods.
There must be strong regulations established with sound ethical considerations before new technology is introduced and begins to spiral out of control. Science and technology themselves are lifeless, but their roles and impacts depend on the people and institutions that control them.
CONCLUSION
Potential 6G technologies and challenges are summarized in Figure 3. Whatever 5G brings, it is time for scholars to take it a step further. This article attempts to glimpse beyond 5G, at AI-defined wireless communication systems.
Figure 3 Potential 6G technologies and challenges.
References
- K. David and H. Berndt, "6G Vision and Requirements: Is There Any Need for Beyond 5G?" IEEE Vehicular Technology Magazine, Vol. 13, No. 3, September 2018, pp. 72-80.
- M. Katz, M. Matinmikko-Blue and M. Latva-Aho, "6Genesis Flagship Program: Building the Bridges Towards 6G-Enabled Wireless Smart Society and Ecosystem," IEEE 10th Latin-American Conference on Communications, November 2018.
- F. Boccardi, R. W. Heath, A. Lozano and T. L. Marzetta "Five Disruptive Technology Directions for 5G," IEEE Communications Magazine, Vol. 52, No. 2, February 2014, pp. 74-80.
- Plattform Industrie 4.0. (2016), Aspects of the Research Roadmap in Application Scenarios, Federal Ministry for Economic Affairs and Energy. Berlin, Germany. Web. https://www.plattform-i40.de/I40/Redaktion/EN/Downloads/Publikation/aspects-of-the-research-roadmap.pdf%3F__blob%3DpublicationFile%26v%3D10
- R. Zhang and C. K. Ho, "MIMO Broadcasting for Simultaneous Wireless Information and Power Transfer," IEEE Transactions on Wireless Communications, Vol. 12, No. 5, May 2013, pp. 1989-2001.
- O. P. Gandhi, G. Lazzi and C. M. Furse, "Electromagnetic Absorption in the Human Head and Neck for Mobile Telephones at 835 and 1900 MHz," IEEE Transactions on Microwave Theory and Techniques, Vol. 44, No. 10, October 1996, pp. 1884-1897.
- M. Yao, M. Sohul, V. Marojevic and J. H. Reed, "Artificial Intelligence Defined 5G Radio Access Networks," IEEE Communications Magazine, Vol. 57, No. 3, March 2019, pp. 14-20.
