[introduction] artificial intelligence (AI) revolution has come. Wait to be released publicly of application as ChatGPT, people is able to use deepness nerve network and machine study (ML) force and latent capacity gain personal experience. ChatGPT is a linguistic model, this model used the huge text data that comes from Internet and book to undertake training, can make the text version of kind of presumable person compose. The used ideal system to reveal artificial intelligence dominant position of this kind of type. It can pass much training data to optimize the output below complex setting ceaselessly.

Wireless network is complex substantially, can generate much data, and the introducing as every generation new technology, its complexity also is increasing ceaselessly. These character make artificial intelligence makes the ideal tool that optimizes wireless network.

01AI applies mediumly in 5G network

As the maturity of 5G technology, AI and ML already by 3GPP (the 3rd acting partner plans) introduce research, 3GPP is the internationalization standard organization that establishs beehive technology standard. Considering to apply artificial intelligence to undertake improvement to the interface in sky at present, include a network energy-saving, load is balanced optimize with mobile sex etc. As a result of the interface in sky use potentially the exemple is very much, chose only in the 3GPP R18 that is about to release so among them a small subclass undertakes study, cover channel condition news (CSI) feedback, beam management and fixed position. Those who need an attention is, 3GPP did not develop artificial intelligence / machine study model. Contrary, it tries to found general framework and evaluate a method, so that artificial intelligence / deploy of machine study model arrives in sky in the different function of interface [1] .

Besides the interface in 3GPP and sky, how is O-RAN alliance using artificial intelligence in exploration / the machine learns improvement to the network weaves and manage. For example, the framework of O-RAN alliance has a distinctive function, this framework is called RAN intelligence controller (RIC) , basically use at assisting artificial intelligence and machine study to optimize different use setting. RIC can manage close real time to apply already (XApps) , also can manage be not real time application (RApps) . The XApps that is used at improving spectrum efficiency and energy efficiency and intelligence of labour of benefit choose and employ persons undertake the network weaves the RApps with management has existed at present. As the development of O-RAN ecosystem and maturity, will more XApps/rApps and the application that use the artificial intelligence that is based on RIC and machine study to optimize will appear.

Graph 1: ORAN network

02 6G network gives birth to AI technology formerly

Although 6G is in start level, but what can decide is, artificial intelligence / machine study will become future wireless communication system the main component of each respects. In network level, although do not have official definition, but ” AI is born formerly ” this one term has been used extensively inside course of study. One of way that observe these AI give birth to a network formerly are a basis RAN (wireless accept a web) current fictitious change technology and depolymerization trend to conclude on graph (graph 1) . Every area piece in the network includes artificial intelligence possibly / machine study model, these models are between different supplier and application may somewhat difference (graph 2) .

Graph 2: ORAN 6G network

Former unripe network also can use AI to point to call run raw raw artificial intelligence / the network that the machine learns a model and compose builds. Ask the design technological process below reference (graph 3) . In traditional 5G network, the interface in sky is comprised by different share, every part all is undertaken designing by the mankind. In 5G-Advanced network, every part will use machine study technology to optimize specific function. In 6G network, may use deepness nerve network to design the interface in whole sky by artificial intelligence.

Graph 3: Develop from what give birth to a network formerly to AI with AI union [2]

03 artificial intelligence / machine study is optimized

Draw lessons from artificial intelligence / the idea that machine study can use at improvement to the network weaves and manage, 6G sends a hope to optimize a challenge at using artificial intelligence and machine study to solve. For example, artificial intelligence can move to the circumstance is opened and shut package according to real time, in order to reduce the power comsumption of whole network. Nowadays, XApps and RApps are mixed through open close the high cost such as the power amplifier that is in non operating state can package is in base station level achieved this one goal. However, artificial intelligence solves the computational problem that has challenge sex and the ability that analyse huge data quickly, it is for us inside bigger range and even whole town perhaps optimized network function to offer a possibility inside countrywide limits. Small time can be compared in use frequency paragraph shut whole base station, also can undertake be configuringed afresh to the village, the real time demand that with green the means of low carbon, energy-saving environmental protection makes exhaust may little resource will satisfy an user. Still cannot configure afresh with this kind of means at present base the network of station and whole town, new configuration and test configure to the network any changing need a few days normally or a few weeks time. Nevertheless, the development perspective of technology of different artificial intelligence is very wide, what they remain infrastructure provider is principal think element.

04 summary

The artificial intelligence application in wireless network won’t when the ability when 6G network appears begins. Whole ecosystem is having active research, in order to develop new model, and these models compositive in the existent wireless communication system with future. However, these models are remained new roll out, need rigorous to its sex and dependability to undertake assessment. Train artificial intelligence model appropriately on different data set, quantify them to be opposite the improvement of traditional technique, the module that is artificial intelligence drive defines new test method, these are the as the adoption of new technology crucial step that must take. As artificial intelligence model and the maturity that test method and technology, without doubt, artificial intelligence will be in prospective 5-10 changes wireless communication industry thoroughly inside year.

Reference material:

[1]https://arxiv.org/ftp/arxiv/papers/2201/2201.01358.pdf

[2] [Hoy21] J. Hoydis Et Al, “Toward A 6G AINative Air Interface” , IEEE Comm. Magazine, may 2021.

Avoid duty statement: The article is reprint an article, reprint this article purpose to depend on passing more information, the person that copyright puts in original work ‘s charge is all. If involve work copyright issue,article place uses video, picture, written language, contact please small make up undertake handling.