Zhejiang University Chinese

What's on at ISEE

Current Position : homepage  What's on at ISEE  NEWS

Prof. ZHANG Chaoyang: Wireless AI May Change the Architecture of Classic Communication System in 6G Network

Date:2020-06-30

At the beginning of 6G network design, the industry wants to design AI into neural network of 6G. ZHANG Chaoyang, Professor of Zhejiang University, said at recent seminars about 6G, that wireless AI may change the modular architecture and communication paradigm of classic communication systems, and even realize 6G intelligent air-interface.

 

Large-scale and complex communication system in future

ZHANG said that the communication system in future would be a large-scale and complex system, and there would also be super long-distance connections between air, sky, earth and sea. The application scenarios are so complex and diverse, in which people, machines and objects exist, together with the increasing dynamic characteristics of the network.

The classical communication paradigms are often based on limited local scenarios, which are highly dependent on channel models, said ZHANG, so it’s a must to make complex measurements for different channel models, overcoming some obvious shortcomings of the classical communication paradigm like the missing details of deterministic scene.

 

The principle framework of wireless AI has been initially established

Artificial intelligence introduces an avenue for communication architecture and paradigm shift. ZNANG believes that through the distributed data at edge and cloud, as well as the mining and utilization large state space, the perception and learning ability of the network will be greatly improved, thus the global optimization will be realized, and the formation of network intelligence will be promoted to provide highly intelligent services.

At present, wireless AI has been highly valued and concerned by research institutions at home and abroad. The theoretical framework and system architecture of wireless AI have been initially established, while AI optimization, migration, compression and other theories are developing rapidly. The mechanism and corresponding performance limit of AI enhancing wireless communication system performance need to be further explored.

Lots of universities like MIT, UIUC, GaTech and Zhejiang University have carried out relevant exploration, and proposed various algorithms. IEEE communication society has launched the machine learning based communication emerging technology research program, and the EU Horizon2020 has also started up the intelligent network and services 6G research project.

In recent years, the international discussion of AI for 5g and 6G has been very active. Some experts have proposed that 6G network will develop from software to intelligence, with the ability of environment perception, resource mining, distributed learning, and even realize 6G intelligent air interface. Some have proposed a machine learning architecture for the Internet of things.

Domestic scholars also propose to use wireless AI to do in-depth mining of spectrum, energy, cache, computing and other resources to enhance the autonomous ability of wireless network, which is suitable for dynamic deployment scenarios such as Internet of things and Internet of vehicles.

ZHANG said that wireless AI will realize the optimization design of end-to-end network system, and can also make full use of the transmission samples of the network, turning data and computing power into a part of resources to further improve resource efficiency, and deeping the integration of communication, computing and storage. In the future, it will take the network as the platform, the user as the center, the data as the raw material, and the computing power as the kinetic energy, so that the comprehensive service capacity of the network will be significantly improved.

 

At the same time, however, wireless AI also faces many basic problems, including architecture, data, model, algorithm and so on. The essence of these problems is the integration of communication and computing and is also the frontier intersection of physics and mathematics in the information field, which needs in-depth research and exploration urgently.


Contact Us