Durham researcher uses AI to aim high with drone safety

Autonomous aircraft are beginning to show promise in a host of applications, from delivering mail between remote Scottish islands to transporting medical supplies between healthcare facilities across the Solent. But if the use of drones was to significantly increase in more city or suburban locations, so too would the risk of collision with buildings. 

To help investigate the threat such craft could pose in urban environments – and come up with ways to mitigate the risks – academic Dr Hubert Shum is developing technology that promises to keep a closer watch on drones in flight. His study received £30,000 in funding from the Transport Research and Innovation Grants programme, delivered by Connected Places Catapult for the Department for Transport. 

“Drones are very powerful tools for transportation but they can also be very dangerous and pose a risk to infrastructure, especially if they are misused,” Hubert explains. “My project is about improving the detection and tracking of drones in order to protect vulnerable buildings.” 

His work involves developing computer vision systems using Artificial Intelligence and machine learning to analyse images and real-time videos of drones in flight, in order to spot when an airborne craft might collide with fixed infrastructure, and allow corrective action to be taken. 

“I used my TRIG money to hire two PhD students as part-time researchers to carry out development and research – as part of a job share – and to work with a specialist company on validation and testing,” he says. “I received lots of support from the Catapult in terms of helping me to understand contracts and funding, and several of their events allowed me to meet with likeminded people. 

“Overall, the experience was excellent and the project was a big help; it kick-started my research and development in this field and helped me to develop my reputation in this area.” 

Dr Hubert Shum, Associate Professor of Computer Science at Durham University

In a further phase of AI study, Hubert is investigating how the technology could be brought together with security cameras to spot when problematic human behaviours are able to escalate in public environments. He has also developed AI systems for use in driverless vehicles, and is also looking to see how the technology could be used to explore satellite imagery to monitor planetary changes.

Hubert grew up in Hong Kong and came to the UK two decades ago to study his PhD in computer science at the University of Edinburgh. He later joined Northumbria University before moving to Durham in 2019. He says computer science is the “perfect place where you can marry the beauty of art and the beauty of science, put them together and develop something very interesting”. 

One important area of focus for him now is training PhD students in AI research so they are better able to “contribute to society when they graduate”. 

Hubert remarks that AI is “probably one of the fastest developing areas of scientific research” and adds that it presents “many good opportunities to solve problems now that could not previously be done.” 

Keep an eye out for TRIG updates in the upcoming months. You can sign up to our SME Network here or our Academic Network to be kept up to date on the latest news.