AI 如何缓解无线电频谱的稀缺竞争

AI is helping ease competition for scarce space on the airwaves

Every time you make a phone call, stream a video or connect to Wi-Fi, your device is sending and receiving invisible signals through the air. So are your baby monitor, garage door opener, your GPS, the weather-monitoring satellites and the radars keeping planes safely apart. All of these rely on the same invisible resource: the radio spectrum.

The demand for wireless data – driven by smartphones, video streaming and billions of connected devices – keeps growing, making available spectrum scarce. But it doesn’t have to remain that way. As an electrical engineer who studies wireless communications, I’m part of a field of research working to address the radio spectrum supply crunch.

The solution engineers and regulators have been developing is dynamic spectrum sharing. The core idea is straightforward: Instead of reserving a portion of spectrum exclusively for one user forever, allow multiple users to share the same frequencies – intelligently, and with clear rules to keep them from stepping on each other’s signals.

Radio spectrum is at the low end of the electromagnetic spectrum, with frequencies below microwaves, light and X-rays. You can think of it as a vast collection of invisible highways, running at different frequencies, carrying signals through the air. Television, air traffic control, cellphones, remote sensing of distant stars – they all travel on different parts of this same invisible road network.

These communication highways span an enormous range, from very low frequencies used for submarine communication all the way up beyond radio waves to the frequencies of visible light. Higher frequencies carry signals over shorter distances but can move more data with wider highways. Because there is only so much spectrum to go around as a matter of fundamental physics, governments have historically managed it by issuing exclusive licenses – dedicating certain highways to specific users and telling everyone else to stay off.

This system worked well enough for decades, especially because the highway network was sufficient for the demand. Studies measuring actual spectrum use over many years have found something surprising: Large stretches of licensed spectrum sit idle much of the time. For example, a military radar system may only operate in certain locations on certain days.

Burgeoning demand for radio spectrum has left some of these roads clogged with traffic jams, while others nearby sit almost completely empty.

Radio communications occupy a small portion of the electromagnetic spectrum. NASA

Traffic control for radio

Making dynamic spectrum sharing work requires radios that are more flexible than the fixed-frequency devices of the past.

Today’s smartphones and wireless devices are increasingly built on software-defined radio technology, meaning the frequencies they operate on can be adjusted by software rather than being locked into hardware. Think of it like a car that can switch lanes automatically rather than being stuck in one lane forever. Early researchers called the vision behind this “cognitive radio” – the idea of a radio that could sense its environment and adapt accordingly.

This flexibility opens the door to a layered approach to spectrum sharing. At the top of the hierarchy sit incumbent primary users – those who hold existing rights to a frequency, such as the military or licensed broadcasters. Below them, other secondary users can access the same spectrum when and where the incumbents aren’t using it, stepping aside automatically the moment an incumbent returns.

A more everyday version of this kind of sharing already exists in the Wi-Fi frequencies you use at home. Multiple devices – your laptop, your phone, your smart TV – share the same spectrum without a central authority assigning each one a dedicated slot. They do this by briefly sensing whether the channel is busy before transmitting, and by automatically adjusting their speed based on how crowded things get. The protocol resembles drivers at a four-way stop – informal rules that mostly work without anyone directing traffic.

A significant recent example of dynamic spectrum sharing in the United States is the Citizens Broadband Radio Service, or CBRS, operating in the 3.5 GHz frequency band. The Federal Communications Commission created a three-tier system. Incumbent users – primarily U.S. Navy radar systems – sit at the top. Businesses and organizations that pay for priority access sit in the middle. And the general public can use whatever is left over.

A central computer system – the Spectrum Access System – acts as a real-time traffic controller, constantly tracking where Navy radar is active and automatically coordinating access for everyone else. When a Navy ship is not using the spectrum in your area, hospitals, factories, warehouses and campuses can use it for their own private wireless networks. When the Navy needs it back, everyone else steps aside within seconds.

The military has priority for using scarce radio spectrum. U.S. Army photo by Cpl. Leonard Beckett

Getting around interference

The biggest technical hurdle in spectrum sharing is interference – what happens when two signals overlap and scramble each other, like two people shouting different things in your ear at the same time. As more devices crowd the airwaves, each with different power levels and different bandwidth needs, keeping signals from colliding becomes increasingly complex.

My lab and other researchers are now turning to artificial intelligence for help. AI can learn from experience and adapt in real time to changing conditions – deciding who transmits when, and at what power level, in ways that older rule-based systems could not. At the same time, new challenges are emerging, such as satellites and ground-based networks increasingly needing to share the same frequencies.

The way radio spectrum is managed may not seem like a kitchen table issue, but it shapes everything from how fast your phone loads a video to whether a factory can automate its floor, to the reliability of a rural school’s internet access. Future wireless technologies, including those that connect self-driving cars and enable remote surgery, will depend on having access to enough spectrum. Dynamic spectrum sharing is how to make room for all of it without dismantling the systems already in place.

The airwaves are finite. But with smarter management, they can do far more than once thought possible.

This article is republished from The Conversation under a Creative Commons license. Read the original article.

The Conversation is an independent source of news and views, sourced from the academic and research community and delivered direct to the public.

AUAI is supported by:

Related posts :

ACM SIGAI Autonomous Agents Award 2027 open for nominations

ACM SIGAI

20 Oct 2026

Nominations are solicited for the 2027 ACM SIGAI Autonomous Agents Research Award.

How AI is helping cities tackle invisible pollution

Horizon

16 Oct 2026

Researchers are designing AI-powered virtual cities to help authorities predict street-level pollution and act where it is needed most.

#AIES2026 social media round-up

AIhub

15 Oct 2026

Find out what participants got up to at the Conference on Artificial Intelligence, Ethics, and Society.

Credit fairness: lending resources without losing out

14 Oct 2026

Studying the allocation of shared resources among agents with time-varying demands.

What’s on the programme at #AIES2026?

Lucy Smith

12 Oct 2026

The ninth conference on AI, ethics, and society is taking place in Malmö from 12-14 October.

Generative AI has changed mathematics forever. Where to from here?

The Conversation

12 Oct 2026

Large language models such as ChatGPT and Claude are creating disruptions in all parts of the mathematical community.

Cybersecurity, evolutionary game theory, and AI safety: an interview with Adeela Bashir

Lucy Smith

09 Oct 2026

In the latest of our interviews with the IJCAI-ECAI 2026 doctoral consortium participants, we learn about how cybersecurity attack and defence behaviours evolve.

Lucy Smith

08 Oct 2026

We collate our interviews with the 2026 cohort of doctoral consortium participants.

Association for the Understanding of Artificial Intelligence