The FlyingToolbox support drone system Loughborough University
Researchers at Westlake University in China have developed a new system that allows drones to exchange tools while flying.
The system, called FlyingToolbox, marks the first time that multi-rotor drones have performed precise mid-air docking and tool transfers.
When one drone flies directly above another, the downward airflow from its propellers creates strong turbulence. This airflow, known as downwash, can reach speeds of over 13 meters per second and disrupt the stability of the drone below.
The challenge has made it nearly impossible for drones to fly in stacked formations or perform precise operations together.
Tackling the downwash problem
The FlyingToolbox system solves this by combining two types of drones. The lower “toolbox drone” carries multiple tools, while the upper “manipulator drone” uses a robotic arm to pick up and return tools mid-flight.
Together they perform tasks much like a surgeon and nurse exchanging instruments during surgery.
To maintain stability, the system uses a combination of real-time airflow prediction, magnetic docking mechanisms, and onboard visual tracking.
A neural network estimates the strength and direction of the downwash, allowing the lower drone to adjust its position automatically. The drones use QR code tracking to align with sub-centimeter precision even in turbulent air.
Precision and reliability
In experiments, the drones achieved an average docking accuracy of less than one centimeter. Tests included 20 consecutive dockings, all completed successfully, even when strong airflows were present. The design also allowed multiple drones to perform multi-stage tool-switching sequences while hovering in close proximity.
The docking mechanism was critical to the success of the system. The manipulator drone’s arm features a cavity with an embedded metal plate, while the toolbox drone has conical magnetic connectors mounted on elastic cords.
When the two align, the magnets automatically lock into place, and the elastic cords absorb minor errors in alignment.
The results show a significant improvement over previous aerial docking systems, which typically achieved accuracies of six to eight centimeters.
The Westlake team’s approach offers a level of control and repeatability that could make drones capable of carrying out more complex collaborative missions in the future.
The FlyingToolbox demonstrates how drones can move beyond simple inspection or filming roles to perform real aerial work.
These flying manipulators could be deployed in environments where sending humans is dangerous, such as disaster zones, high-rise maintenance, or industrial inspections.
While current testing was limited to controlled indoor environments, the researchers plan to adapt the system for outdoor use.
They are also developing robotic arms with more degrees of freedom and improved perception systems to handle more complicated manipulations.
The team believes the progress could transform how drones are used for cooperative tasks.
By enabling them to share tools and divide labor, drones could operate as coordinated aerial workforces capable of executing missions far beyond their individual limits.
The study is published in the journal Nature.
MasterCard
