How Can Cloud Computing Improve UAV Operations?

Unmanned aerial vehicles could play a key role in several industries, but their underlying technologies need more refinement before they reach their full potential. Cloud computing could be the answer.

By Emily Newton

Unmanned aerial vehicles have become essential technologies for a wide range of industries and applications — including agriculture, firefighting, and construction.

In the near future, they may also be an essential tool for major retailers, who hope to use the drones to quickly deliver essential items directly to consumers.

However, despite rapid advances in UAV technology, certain applications of the tech remain out of reach. Real-time data collection and processing with UAVs, for example, often isn’t workable right now.

Cloud computing — and related tech, like edge networking — could make advanced uses of UAVs more effective and more practical.

Current Challenges for UAV Operations

Often, UAVs use radio frequency transmission to communicate with a ground station — receiving commands from a pilot or piloting system and transmitting data from the air. This method of in-flight data transfer, while effective for some purposes, has a few limitations.

With radio frequency transmissions, the user or ground station needs to always be in a certain location near the drone. The method also limits the size of mission areas — fly too far from a ground station, and the UAV could risk losing the signal.

This, along with limited power supplies, makes large mission areas or long flights — for example, inspecting a pipeline for damage, providing information on wildfires to firefighters, or making a long-distance delivery — less practical.

The approach also limits the amount of data that can be transmitted between the drone and the ground station. Many missions, as a result, focus less on real-time data and more on data captured in the air that will be delivered before being processed.

In some cases, the time-to-delivery won’t matter. In others — like in search-and-rescue or wildfire analysis — that delay can have serious consequences.

Real-Time UAV Data With the Cloud

It may be possible to overcome these challenges with cloud computing. If a UAV is able to communicate with the cloud in the air — or with edge network nodes — it could provide near-real-time data to users.

For example, a team of Spanish researchers from the University of Barcelona used a cloud-connected UAV to gather real-time info on olive trees while the drone was in flight. The researchers were able to geolocate and digitally represent the trees within seconds of the drone taking off, without needing long processing times or drone recovery before the delivery of usable information.

While the scope of the project was limited, the researchers also suggest the tech could make a number of applications possible — like cloud-and-UAV-powered search and rescue, fire detection, and stock breeding.

The use of the cloud also solves a major problem of using advanced data processing or analysis techniques with on-drone hardware.

Data processing, a naturally energy-intensive process, draws from the same power storage that the drone uses to fly. The more you demand, the less time the drone can stay in the air.

This makes it challenging for researchers to take advantage of cutting-edge AI tech, like computer vision algorithms, that can help drones identify objects on the ground while in the air.

This use of computer vision is a great way to make current applications of drone technology more effective. For example, drone site inspection is a popular application of UAV technology. A site-inspection drone could use a computer vision algorithm to identify structural weaknesses or potential safety risks while in the air, allowing a pilot to change course for a closer look.

By offloading computer vision work or similar data processing to the cloud or an edge node, you can significantly cut down on the power a drone uses in-flight.

One research team that experimented with the use of edge nodes and UAVs — through what the researchers called an intelligent Task Offloading Algorithm (or iTOA) — also found that it’s possible to reduce latency using the cloud.

For UAV applications that generate real-time data, this lower latency could be essential.

UAVs as Cloud Infrastructure

Some researchers have also proposed making UAVs and similar technologies part of the cloud infrastructure itself, by using internet-connected UAVs as mobile edge nodes.

These nodes could provide mobile compute resources for a variety of edge devices — especially IoT devices in rural areas, where latency and poor internet infrastructure may make conventional cloud approaches less workable.

For example, some researchers have started to experiment with smart flood detection using IoT devices that track hydrological and meteorological conditions — like humidity or the level of an area river.

Data from these devices, when processed with an AI algorithm, can help predict floods further in advance — giving area officials more time to stage a response or announce evacuations.

However, IoT devices in rural areas typically can’t have wired internet connections and may deal with weak-to-no cellular network signal strength. This often means data will be patchy, unreliable, or impossible to collect.

Using UAVs as a mobile, adjustable edge network could help improve latency without needing to commit significant permanent resources to a traditional edge node.

Using the Cloud to Transform UAV Usage

Cloud computing has the potential to reshape how businesses and researchers use drones.

With the cloud and related tech, like edge networking, it’s possible to process data that drones collect while they’re in the air.

This could have major implications for several drone uses — like search and rescue, agricultural analysis, and flood prediction.

##

About the Author

Emily Newton is the Editor-in-Chief of Revolutionized, where she covers industrial, engineering, and science topics.