Drones on the Edge

In the fifth installment of the ongoing series of webinars from The Global UTM Association (GUTMA), drone services on the edge of the network were considered. Four speakers from different regions and functions addressed this topic. There were organisations from Australia, the US and Latvia – AiRXOS, LMT, Unleash and Airmap.

Focusing on the linkage to telecommunications, LMT shared some interesting concepts. The Latvian mobile network operator described their strategy for moving from drones to phones. Focusing on the need for linkage to the cellular network in order to support beyond visual line of sight (BVLOS) flights, they cite a measurement of the cellular network at already being able to support 97.6% of remote drone management functionality.

There are two topics related to UTM that are currently being addressed. First is a project focused on a mobile network-enabled UTM app.  The mobile (cellular) network provides benefits to such an application that are intrinsic in the network infrastructure.  The use of licensed spectrum, network security and connectivity across the nation, provide a higher level of performance to public and private sectors working with drone fleets.  The concept for the app is to provide:

  • safe LLA (low level altitude) communication network for unmanned aircraft systems (UAS)
  • population density information
  • UAS tracking
  • a detect and avoid system
  • real-time communication throughout the flight (BVLOS).

All UAS that are used for commercial and recreational purposes would be using the app in order to file their desired flight plan to be safely integrated into manned airspace.

The second project incorporates artificial intelligence (AI) into the drone topic to provide an AI-powered rescue drone.  They are working in collaboration with Riga Technical University to develop solutions in support of search and rescue teams.  The concept is that drones can gather a multitude of information through airborne sensors to aid in the rescue effort.   AI would play a role by helping to recognise humans located in remote terrain via information gathered from an infared camera mounted on the drone, allowing for a much wider search in a shorter period of time.  Also reducing costs that would be incurred by sending search teams into remote areas.

Another interesting presenter was Unleash. Ever wondered what to do with all of the information gathered by a drone in order to easily consolidate and analyse it? This is where Unleash comes in, with an AI platform for connected field drones with live feeds.  They talk about making intelligent remote vision possible via their cloud-connected platform that also provides storage of the content in addition to AI-enabled analytics.

The AI analytics can:

  • convert unstructured video into insights with machine learning
  • enable faster decisions from notifications powered by custom AI algorithms
  • track and monitor objects, live streamed and viewed anywhere
  • automatically extract clips or highlights with AI
  • visualise data trends based on historical data.

As more and more companies enter this space with competing solutions, it will be interesting to see what happens when they all have to interoperate to share data.  It seems some will be better positioned than others to succeed in the long term if they have started with an open platform from the beginning.


ATM had the opportunity to ask the panelists for their views on a few additional questions outside of those presented during the webinar.

Q – If drones are used to gather data – what types of data would you see allowable under strict privacy rules such as GDPR?

A – The answers provided followed two opposite trains of thought.  On the one side, the view was that regulation would be ‘modified’, maybe too strong a word, to allow for less privacy in order to complete a task.  In other words, if the drone was collecting data about a specific task and outside the focus area, information was gathered not pertaining to this task, wouldn’t be considered a problem – such as a license plate of a passing car.  I think that this is very dependent on where you live if this would be acceptable or not.

Another answer focused on a real-world example related to the loosening of restrictions due to COVID-19.  Using drones to track movement in areas where rules are relaxing, but maintaining anonymised information to obfuscate any personal data – face, car license plates, etc.  This allows for group sizes and movements to be tracked and addressed if outside of the current regulations without sacrificing any privacy.

Q – How would the drones interoperate with the mobile network and where do you see the data being stored?

A – The answers were quite consistent here that 5G is really necessary for this to be successful as data with lower latency and higher bandwidth requirements – live streams – is gathered.  The mobile network operators are well equipped to store sensitive data and have been doing so for years.  Their secure cloud infrastructures provide a ready-made solution for this.

Q – How can AI utilise information gathered from drones to help companies with drone fleets optimise their services?

A – AI enables drones to ‘perceive’ their surroundings.  What this means is a complete transformation of how enterprises conduct visual inspections, monitor and report on the findings.  In this situation, AI can be trained to be more accurate than humans, when observing from a distance for any potential faults requiring maintenance, for example.  The AI would not replace the human, but augment the human with more precise information gathering.