Control system to help multiple drones work together

Many of the parcel delivery drones of the future are expected to handle parcels weighing up to five kilograms, which would allow small, standardised UAVs to take over a large percentage of the deliveries currently made by ground vehicles. But will heavier packages be left to slower traditional truck and van deliveries?

A team of researchers at the Georgia Institute of Technology has developed a modular solution to handle larger packages without the need for a complex fleet of drones of different sizes. By allowing teams of small drones to lift objects in cooperation using an adaptive control algorithm, it could be possible to deliver a wide range of packages using a combination of several standard-sized vehicles.

Beyond simplifying the drone fleet, the work could enable more robust drone operations and reduce the noise and safety concerns associated with flying large autonomous UAVs in populated areas. In addition to commercial package delivery, the system could also be used by the military to deliver supplies to small groups of soldiers in the field.

“A delivery truck could carry up to a dozen drones in the back, and depending on how heavy a particular shipment is, up to six drones could be used to deliver the package. This would give flexibility in the weight of packages that could be delivered and eliminate the need to build and maintain multiple delivery drones of different sizes,” said Jonathan Rogers, research scientist for avionics integration at Lockheed Martin.

The idea is based on making it easy for users to fly multi-UAVs together. The difficult questions are solved by on-board intelligence, rather than relying on a human to accurately measure the weight of the package, its centre of gravity and the position of the drones relative to each other. The aim is to make it so simple that even a parcel delivery driver can operate the system without any problems.

The biggest challenge in the development process is that most drone swarms involve autonomously manoeuvring vehicles that are not linked but fly in formation. In this case, the unique dynamics of a particular vehicle are not limited by what other drones are doing. However, the drones attached to the package are subject to the forces exerted by the other vehicles in addition to the weight of the load being carried.

The team of drones would autonomously attach to the docking device attached to the package, using an infrared guidance system that eliminates the need for humans to secure the vehicles. This could come in handy for drones flying to pick up packages returned by a customer. By knowing how much thrust they are generating and at what altitude they are, drone crews could even estimate the weight of the package they are picking up.

In a demonstration put together at the university, four small quadrotor drones work together to lift a 60x60x60 centimetre box weighing 6 kilograms. The control algorithm is not limited to four vehicles, it can handle as many as can fit around the package.

Georgia Tech researchers believe that the core technologies are already in place and, with the right investment, a system that could deliver packages using multiple drones could be deployed within five years. This is not so much a technical challenge as a question of regulation and social acceptance.