The system combines a centralized kinodynamic motion planner with onboard controllers, enabling real-time coordination of the entire multi-quadrotor-load system. This allowed the system to achieve speeds exceeding 5 m/second and accelerations greater than 8 m/s2, perform maneuvers like flying through narrow passages, and remain robust to disturbances, far surpassing current methods.
Overcoming the Limitations of Multi-Drone Payload Transport
The use of drones to transport goods has grown rapidly, but a key limitation remains: the relatively low payload capacity of a single drone. This constraint significantly reduces their usefulness in critical applications, whether it's delivering heavy construction materials to remote locations, moving large harvests from agricultural fields, or conducting rescue operations that require transporting victims or bulky equipment quickly and safely.
One proposed solution, attaching multiple drones to a single payload using cables, seems straightforward in theory but introduces major control challenges. When drones are connected this way, they become dynamically coupled: the movement of one drone directly affects the others and the suspended load. This creates a complex, swinging system that’s difficult to manage, especially under external disturbances like wind.
Traditional control algorithms aren’t equipped to handle this level of interaction; they tend to be slow and rigid, limiting current multi-lift systems to cautious, low-speed operations. As a result, these setups fall short in time-sensitive scenarios that demand speed, precision, and agility.
This research tackles those challenges head-on by moving away from conventional, layered control strategies and adopting a unified "whole-body" planning approach.
Rather than treating each drone and the payload as separate components, the new algorithm plans their movements as one fully integrated system in real time.
By removing the need to isolate the drone and load dynamics on different timescales, the framework accounts for all physical coupling and constraints at once. This allows the system to anticipate and actively counteract dynamic forces and load swings, instead of simply reacting after the fact. The result is a dramatic improvement in speed, responsiveness, and stability - making multi-drone transport viable for demanding, real-world missions.
A Unified, Whole-Body Planning Framework for Agility and Control
The core innovation of the research lies in its fundamental departure from traditional control methods. Previous systems often used a cascaded approach, where the trajectory of the load and the positions of the drones were calculated in separate steps. This method is inherently slow and fails to fully account for the real-time, dynamic interplay between the quadrotors and the suspended object. It is this very limitation that has historically capped the speed and acceleration of multi-drone lifting systems, making them cautious and sluggish.
The new algorithm introduces a trajectory-based framework that tackles the whole-body kinodynamic motion planning problem. In essence, this means the system plans the motion of the entire assembly as a single, cohesive unit.
It calculates optimal trajectories in real time (online), explicitly considering the complex physical forces and constraints linking the drones to the payload.
The framework consists of two layers: a centralized online kinodynamic motion planner that runs at 10 Hz and onboard incremental nonlinear dynamic inversion (INDI) controllers operating at higher frequency to track these trajectories. This holistic approach allows the drones to anticipate and compensate for each other's movements and the swinging of the load, rather than just reacting to it after it happens.
The planned trajectory is then provided to the quadrotors in a receding-horizon fashion, meaning the system constantly updates its plan based on the current state, allowing for dynamic adjustments mid-flight. This results in a level of agility previously thought impossible. The team's experiments show the system can achieve over eight times greater acceleration than state-of-the-art methods.
Real-World Testing Demonstrates Robustness and Precision
To validate their algorithm, the team developed a complete system using three custom-built quadrotors, putting it through a demanding series of tests in a controlled lab environment. These trials were carefully designed to mimic the unpredictability of real-world conditions.
A fourth quadrotor was later added for extended robustness testing. The researchers introduced obstacles requiring precise navigation, simulated wind disturbances with high-powered fans, and, most notably, used a dynamically shifting payload, such as a basketball, to challenge the system’s ability to handle significant mass and inertia mismatches.
The system passed all tests convincingly. A standout demonstration involved executing high-speed maneuvers through narrow passages - highlighting its agility and precision. The onboard controller actively monitored and adjusted for cable tension, eliminating the need for sensors on the payload itself. This makes the setup far more flexible, as virtually any object can be lifted without modification.
The system also demonstrated strong resilience, maintaining stable performance with over 40 % variation in payload mass and in wind speeds up to 5 m/second. These results underscore its potential for reliable deployment in the harsh, unpredictable environments where it's needed most.
Advancing Cooperative Aerial Robotics Toward Real-World Deployment
In conclusion, the development of this fast, flexible, and robust algorithm marks a significant leap forward for cooperative robotic systems. By enabling teams of drones to transport heavy loads with unprecedented speed and agility, the technology shatters previous performance barriers.
Its two-layer architecture, combining predictive planning and adaptive onboard control, demonstrates how multi-robot coordination can be both agile and stable. Its ability to function effectively without payload sensors and in windy conditions strongly enhances its real-world applicability. While current testing relies on indoor motion capture, the foundation is now laid for future development aimed at outdoor deployment.
Journal Reference
Sun, S., Wang, X., Sanalitro, D., Franchi, A., Tognon, M., & Alonso-Mora, J. (2025). Agile and cooperative aerial manipulation of a cable-suspended load. Science Robotics, 10(107). DOI:10.1126/scirobotics.adu8015. https://www.science.org/doi/10.1126/scirobotics.adu8015
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.