Last year, while working on a UAV swarm project, my understanding of 5G modules was still limited to the parameter level: high bandwidth, low latency, and impressive specifications. It was not until we later worked on two real projects - a campus low-altitude management system and an autonomous tunnel inspection system - that the limitations of traditional communication solutions in complex scenarios gradually became clear.

In this article, I want to start from that experience and explain why I have become increasingly convinced that 5G will become a key solution as UAV communication moves toward large-scale application.

01

The Capability Boundaries of Two Traditional Solutions

The most common UAV communication solutions on the market essentially fall into two categories: video/data radio links and wired fiber optics.

Video/Data Radio Links

This is the solution we use most often. An integrated video transmission and data telemetry radio can carry MAVLink telemetry and HD video. Within visual line of sight, latency can reach 20-50 ms. It is stable, direct, and relatively low cost.

The problem is also very clear: it depends on line-of-sight propagation of electromagnetic waves. Once the distance exceeds 1.5-2 km, or if buildings, mountains, or tree canopies block the path, the signal drops sharply. The MAVLink heartbeat interval stretches from 1 second to 5-8 seconds, and the aircraft attitude indicator on the ground station starts behaving erratically. In practice, the aircraft is already close to a semi-out-of-control state.Video/data radio links need to be evaluated in specific scenarios. For line-of-sight, medium- and short-range, high-bandwidth tasks, they remain a very common communication solution.

Take the LQ-10 long-range video/data transmission module as an example. It supports integrated video transmission, data transmission, and star networking. In broadband mode, the transmission rate can reach 100 Mbps at 1 km and 80 Mbps at 1.5 km. In narrowband mode, the ground-to-air distance can reach 7 km. On UAV platforms such as the P600, it is relatively simple to deploy for research debugging, teaching and training, short-range inspection, and HD video backhaul, and the link architecture is also easy to understand.

However, in beyond-visual-line-of-sight operations, complex obstruction, multi-aircraft online operation, and compliance reporting scenarios, relying solely on video/data radio links becomes limiting. Wired Fiber Optics

In a tunnel inspection project, we tried a fiber-tethered solution:

The UAV flew while dragging a fiber-optic cable. The bandwidth was indeed sufficient, and the latency was almost zero, but the core problem was that the aircraft was tied down by a cable. It could only fly along the fiber path, and friction at bends continuously consumed endurance. In multi-aircraft collaboration, each aircraft would also need its own independent fiber link. The more aircraft there are, the less feasible the system becomes. Fiber is better suited to "maximum reliability on a fixed route," while low-altitude management and inspection need "flexible coverage + always-on connectivity."

02

Two Key Realizations That Reshape Selection Logic

5G Is Already-Built Urban Infrastructure

According to data from the Ministry of Industry and Information Technology, by the end of 2024 China had 4.251 million 5G base stations, achieving "gigabit access in every county and 5G in every township." Beijing Metro is promoting 5G upgrades on 11 existing lines, covering 137 stations and 202 km of tunnels, with leaky coaxial cables laid along the full tunnel routes. Shanghai has also completed full 5G coverage on 6 lines, 34 stations, and 61 km of tunnels.

Even in fully enclosed, metal, tubular underground spaces, 5G signals can maintain continuous end-to-end coverage. For those reinforced-concrete buildings above ground, there is no need to force the signal to "punch through" them; another base station is often only a few hundred meters away.

UAV Communication Also Needs to Become Standardized

Manned aviation has long formed a relatively unified communication system. Whether Boeing, Airbus, or COMAC C919 aircraft are involved, they all use aeronautical mobile communication frequency bands allocated by ICAO, the International Civil Aviation Organization. Communication and dispatching between aircraft, and between aircraft and the ground, can be carried out under a unified standard.

The consensus behind this is simple: aviation communication infrastructure must be standardized, universal, and interoperable.

The UAV industry is moving toward a similar stage. In the past, when a single UAV operated in a local scenario, using each manufacturer's own video/data link was not a major problem. But when a low-altitude management system needs to connect multiple UAVs, multiple brands, and different mission types at the same time, the limitations of private links become obvious.

For example, if one P600 uses a 900 MHz radio from Manufacturer A, while another M300 uses DJI's proprietary link, how can they be uniformly identified, managed, and dispatched on the same low-altitude management platform?

More importantly, this is no longer just a matter of convenience. Regulations are also pushing UAV communication toward online operation and standardization.

The Interim Regulations on the Flight Management of Unmanned Aircraft, effective January 1, 2024, require medium and large UAVs to report identification information and flight data in real time through the UOM civil unmanned aircraft integrated management platform. In March 2026, the Civil Aviation Administration of China further issued Announcement No. 5, extending network-based operation identification - that is, real-time position reporting - to all civil UAVs, with implementation beginning on May 1, 2026.

This means the role of the UAV communication link is changing. It no longer only carries flight control and video transmission tasks; it is also becoming part of compliant operation.

Future UAV communication must not only support stable flight, but also keep the aircraft continuously online, manageable, and connected to regulatory systems. Can video/data radios support real-time position reporting? Yes, but only if the link remains stable at all times. Once the aircraft flies beyond visual line of sight, or if buildings, mountains, or other obstacles appear in the middle, communication may be interrupted. Regulators require continuous online status and real-time reporting, but the link carries the risk of dropping at any time. That becomes an obvious weakness in compliant operation.

This is also why 5G has more advantages in low-altitude management scenarios. Relying on the coverage density, bandwidth capability, and connection scale of operator networks, 5G can simultaneously carry flight control, video backhaul, and UOM compliance reporting, keeping UAVs online and manageable across a larger range.

Horizontal comparison of three communication solutions 03

Review of Two Real-World Test Scenarios

Within visual line of sight, we can use the LQ-10 to complete high-bandwidth video transmission and stable data telemetry. Once the scenario becomes campus-level, tunnel-level, or beyond-visual-line-of-sight dispatching, the value of a 5G gateway is amplified. In our deployment, we used a 5G SD-WAN gateway optimized specifically for UAV clusters. It does not solve only a single video transmission problem; it places the ground station, aircraft, RTK, and business platform on the same network. Its core capabilities are as follows:

· 5G virtual LAN networking: the ground station and all aircraft are on the same LAN; if ping works, the link works.

· RTK differential correction with one-to-many transmission: the base-station serial port connects to the gateway, the rover serial port connects to the gateway, and transparent transmission starts automatically after power-on.

· Ready out of the box: no VPN configuration and no MAVLink routing hassle; networking changes from a software-engineering task into hardware wiring. 

Scenario 1: Campus Low-Altitude Management System

A university campus covers about 1.2 km2, with teaching buildings, dormitories, and libraries densely distributed. On a normal day, 5-8 UAVs are online at the same time, performing inspection, mapping, and logistics delivery. The platform needs to perceive the position and status of every aircraft and dispatch them uniformly.

Core Challenges

Building obstruction: A campus is not an open field. When an aircraft flies to the north side of a teaching building, three reinforced-concrete buildings may stand between the video/data radio and the ground station. RSSI drops directly below -100 dBm, video freezes, and telemetry is interrupted. With a 5G cellular network, base stations cover the areas around each building. In a cellular architecture, "being blocked by a building" basically does not become an obstacle, because a nearby base station can cover the aircraft from another angle.

Complex network environment: Campus student density is extremely high, with tens of thousands of Wi-Fi, Bluetooth, and 4G/5G terminals. The spectrum environment is extremely noisy. Video/data radios operate in open ISM bands and have no QoS guarantee; when interference appears, the link can only endure it. In peak-hour tests - noon and 5 p.m. - the bit error rate of the data link surged to 5%.

After switching to the 5G gateway solution, the authorized 5G spectrum plus network slicing naturally separated the UAV link from students' video-streaming traffic. This is physical-layer QoS, not software-layer "avoidance." 

Scenario 2: Tunnel Inspection

The site was a 3 km underground utility tunnel in an industrial park, with seven bends. The traditional inspection method was manual entry. We tried two solutions:

Fiber-tethered solution: The UAV dragged a fiber-optic cable while flying, with video transmission and flight-control data carried over the wired link. The latency was indeed extremely low, and the signal was indeed stable.

Core Challenges

· The UAV could only fly along the fiber path and completely lost maneuverability. At bends, the fiber repeatedly rubbed against the tunnel wall, and the resistance cut endurance by 30%.

· Multi-aircraft collaboration? Each additional aircraft required another independent fiber link. The more aircraft were added, the less feasible the approach became.

5G solution: Leaky coaxial cable also needs to be deployed inside the tunnel, just like fiber; both are cables. The difference is:

Advantages

· Leaky coax is shared infrastructure. Once the operator deploys it, every device connected to the 5G network can use it; there is no need to pull a separate cable for each aircraft.

· The aircraft does not need to be "tied down." As long as it remains within leaky-cable coverage, it can fly freely, autonomously avoid obstacles, and switch inspection routes at any time.

Fiber solves the problem of "the signal can be transmitted." 5G solves "the signal can be transmitted + the aircraft can move." The latter is what tunnel inspection truly needs.

04

Conclusion

Looking back at these projects, my understanding of UAV communication selection has changed in two ways.

First, 5G should not be treated merely as a communication module. For beyond-visual-line-of-sight scenarios such as cities and industrial parks, it is more like an already deployed low-altitude communication infrastructure. If UAVs need to stay continuously online, connect uniformly, and backhaul data stably across a larger area, it will ultimately be difficult to bypass 5G networks.

Second, regulations are driving communication-link upgrades. In the past, UAV communication focused more on questions such as "Can it fly far enough?" and "Is the video clear?" Now, real-time position reporting, operation identification, and platform access are becoming basic requirements for compliant operation. UOM real-time reporting is only the beginning. As low-altitude scenarios truly scale up, "flying compliantly and being manageable" will become increasingly important.

Therefore, including 5G in UAV communication selection is not a distant trend. It is a practical requirement that is already becoming visible in project implementation.