One of the biggest challenges facing commercial drone applications such as delivery is the lack of a traffic management system to monitor their movements. Without one, there’s no easy way to integrate drones into the airspace and avoid collisions. Students at Imperial College London have developed what could be the start of a solution.
Aided by Microsoft’s cloud technology and drone traffic management pioneers Altitude Angel, computing students from Imperial College London have shown that over 1,000 drones could fly safely in a one square kilometre area, even with the introduction of manned air traffic.
Read more: Malawi and Unicef open air space for drone deliveries
Repelling mid-air drone collisions
The Imperial College team developed a system to control multiple small, unmanned aircraft at the same time. The project was run and simulated on Azure, Microsoft’s secure cloud computing platform.
The secret to the success of the system is an algorithm that forces drones to be “repelled” if they are about to occupy the same airspace as another object, enabling them to change direction and avoid impact. Taking inspiration from magnetic effects seen every day, this technology could allow multiple drones to navigate a tight space without ever colliding.
When tested against an Altitude Angel simulation, the students proved that many drones could coexist with manned and unmanned aviation, crossing paths safely without the need for human intervention.
Current anti-collision measures in the industry rely on sensors built into drones that detect obstacles. This kind of dynamic avoidance system could be deployed on a large scale, accounting for the movements of many aircraft at the same time.
Read more: Amazon patents inner-city drone delivery towers
“Truly groundbreaking”
The Altitude Angel team, which has experience implementing drone traffic control systems and working with regulators, was impressed with the outcome of the study.
“The students came up with some really innovative designs which complement our existing solution. The proof of concept research undertaken is truly groundbreaking,” said Lawrence Gripper, Altitude Angel’s head of product. “We are already investigating options for further development of these options and their use with our GuardianUTM service.”
“The project work contributes to Altitude Angel being able to explain and demonstrate to stakeholders that the technology for autonomous drone flights is closer than currently believed. It also unveiled potential algorithms that may be adopted during the next development phase of the Altitude Angel platform, accelerating the time to production,” said chief operating officer Chris Forster.
Read more: NASA tests drone traffic control system across the US
Beyond geofencing
Several drone manufacturers, including industry leader DJI, apply geofencing software to products to ensure that pilots can’t fly in restricted areas or above a certain height. While this technology is having positive results, there are still regular reports of near-misses with manned aircraft and drones flying where they shouldn’t.
In late July, a drone forced two flights approaching Lisbon airport in Portugal’s capital to abort their landings, according to reports in the national press. One of them, an Easyjet flight from Paris, was subsequently diverted to Porto, some 200 miles north.
The technology developed by researchers at Imperial College London will be of major interest to commercial operators, particularly those developing beyond visual line-of-sight (BVLOS) capabilities – where the operator has no direct sight of the drone – such as drone delivery.