Body-Motion Control of a Simulated Aerial Swarm from a First-Person View
This paper presents a body-motion interface for first-person-view aerial swarm teleoperation that, through a user study, demonstrated significantly improved navigation efficiency and command stability compared to conventional controllers, albeit at the cost of increased physical demand.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine a team of drones working together to search a collapsed building or map a dense forest. To be truly effective, these machines need to act as a single, cohesive unit, moving through complex three-dimensional spaces while a human operator watches and guides them. The challenge for the human is immense: they must control where the group goes, how fast it moves, how tightly the drones are spaced, and where the camera is looking, all at the same time. Traditionally, this has been done with a handheld remote control, similar to those used for hobbyist aircraft, where the operator uses their thumbs to push sticks in different directions. However, as the tasks become more complex, relying on just two thumbs to manage five different movements simultaneously can feel like trying to solve a puzzle with too many pieces and not enough hands.
Researchers at the École polytechnique fédérale de Lausanne in Switzerland set out to see if using the human body itself could solve this bottleneck. They developed a new way to fly a simulated swarm of fifteen drones using a first-person view, where the operator sees the world through the eyes of one specific drone in the group. Instead of a remote control, they built an interface that translates the operator's natural upper-body movements into commands for the entire swarm. By leaning their torso, moving their hands up and down or apart, and turning their head, the operator could control the swarm's forward motion, side-to-side movement, altitude, the spacing between drones, and the direction the camera was facing. The system was designed to be personalized; before each test, the researchers calibrated the interface to match the specific range of motion and comfortable posture of each individual participant, ensuring that a small lean meant the same thing for everyone.
To test this idea, fourteen volunteers navigated a simulated obstacle course filled with gates and stars to collect. They performed the same task twice: once using the new body-motion interface and once using a standard remote control transmitter. The results showed that the body-motion approach was significantly more efficient. When using their bodies to control the swarm, participants completed the course nearly twenty percent faster than when using the remote. Their path through the air was also more direct, covering less total distance, and they moved with greater smoothness. The data revealed that the body-motion interface allowed operators to adjust multiple commands at once, such as changing the altitude and the spacing between drones simultaneously, whereas the remote control forced them to switch back and forth between different sticks and knobs. This ability to make concurrent adjustments seemed to reduce the need for constant corrections, leading to a more fluid flight path.
Despite these gains in speed and efficiency, the study found that the body-motion method was not a perfect replacement for the remote control in every way. The participants reported that controlling the swarm with their bodies required significantly more physical effort, as they had to hold specific postures and move their limbs continuously. Interestingly, this extra physical demand did not translate into a higher overall feeling of mental stress or difficulty, nor did it affect the quality of the mission. The number of crashes, the number of drones that got separated from the group, and the accuracy of passing through the gates remained the same for both methods. The researchers noted that while the body-motion interface was faster, it did not make the task easier in terms of mental workload, and the participants' preferences for one method over the other were split.
The study suggests that distributing control across the head, torso, and hands can help operators manage complex, multi-dimensional tasks more effectively than a traditional two-stick remote. However, the researchers also pointed out that the physical cost of this efficiency is real, and future designs would need to find ways to reduce the strain on the operator's arms and body. They also observed that the mechanical coupling of the body, where leaning forward might unintentionally lower the hands, could sometimes cause unintended changes in commands, though the participants adapted to this quickly. Ultimately, the work demonstrates that while using the whole body to fly a swarm of robots is physically demanding, it offers a distinct advantage in speed and directness, opening a new path for how humans might interact with groups of machines in the future.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.