Marco Tognon: “In aerial robotics, progress is phenomenal”.

Creating autonomous robots in complex environments

“As soon as I studied engineering, I wanted to concentrate on robotics ”, recalls Marco Tognon. A member of the Rainbow project-team at the Inria center at the University of Rennes, this specialist in aerial robotics was honored last October with the prestigious Toshio Fukuda IROS 2024 prize , awarded to young researchers.

It's an accolade for the man who began working on these issues some ten years ago, at the Max Planck Institute in Tubingen (Germany). Back then, we wanted to develop autonomous robots that could navigate in complex environments,” he recalls. But since technology has evolved so much, my work has changed enormously since then."

The challenge of preventing drone crashes

Regardless of these innovations, the design of robots capable of performing physical tasks autonomously, in different types of environment, remains the guiding principle of Marco Tognon's career. Mainly in the air, his favorite subject.

The challenge is considerable. At a time when drones are becoming widely available and increasingly efficient, it's still tricky to get fully automated machines into the air. “With a ground robot that misses an action, you just press a button and it stops,” illustrates the researcher. But if it flies... It crashes!"

To avoid crashes, robots need to be able to perceive the entire environment around them, thanks to a series of sensors and cameras. But they also need to be able to move, even though they are inherently unstable when in flight. These are challenges that Marco Tognon's research is gradually helping to overcome.

Powerful algorithms to model all simultaneous actions

When flying, a robot constantly exerts a force to keep itself in the air, thanks to the rotation of its blades, for example. Another force is used to propel the robot upwards, downwards, forwards or backwards. All these movements generate a degree of inaccuracy that Marco Tognon seeks to correct. The aim is to ensure that the robot's other simultaneous actions do not conflict with these forces.

And ideally, these machines would need to be relatively powerful to become more agile. In Marco Tognon's view, aerial robots should therefore be equipped with more “muscles” to be able to perform real manipulations. But more “muscle” means more weight, which is a tricky proposition for machines that are supposed to be light enough to fly without difficulty. This balance therefore requires powerful algorithms to model and organize all the actions of aerial robots. A complex challenge at the heart of Marco Tognon's research. It is this work that has led him to receive the Toshio Fukuda Prize today, but not only. The award also recognizes his commitment to the scientific community, as well as his hopes for this rapidly changing world.

Airborne cobots to help humans on the horizon

Progress has been phenomenal,” assures the researcher. When I started, we could barely create small drones rising a few centimetres, pulling cables with very limited force. Now, our robots are capable of flying and opening doors." A source of major advances, his research will not stop there. In addition to current drones, Marco Tognon and the Rainbow team are pursuing another goal: to design aerial cobots. Also known as collaborative robots, these machines are designed to cooperate with humans, working with them to carry out certain tasks.

That's a more distant goal,” explains the researcher. Aerial robots are not yet mature enough to collaborate with humans, at least not without an operator controlling them directly." Nevertheless, Marco Tognon, with the help of several researchers, has already designed a prototype that comes close to this ideal. This is a robot equipped with an arm, developed as part of the FlyHandyBotproject , aimed at creating “work drones”, capable of performing physical labor. Although very promising, this prototype currently only works in the laboratory, and not in real-life conditions surrounded by humans.

Making several aerial robots work together

Another of the researcher's more distant ambitions: to make not just one autonomous aerial robot act, but several at the same time. Here too, the algorithms to be developed must be robust to avoid the risk of accidents. In this type of project, the simplest solution is to have several robots collecting data and transmitting it to a central unit," explains Marco Tognon. This unit then selects the best behavior to adopt."

However, this solution is not applicable to all tasks. And in order to act faster and more efficiently, robots need to be autonomous, while at the same time being “aware” of the environment around them. Better still, they need to transmit the data acquired by their various sensors to their neighbors, so that they too can make the best decision.

What are the possible applications? There are many, as Marco Tognon illustrates: “For example, these robots could help install antennas or change skyscraper windows, or work on high-voltage lines or wind turbines... Dangerous tasks currently performed by humans.”

Ambitious, yes, but not excessive. And with the award he has just received, Marco Tognon and the Rainbow team have an additional asset with which to demonstrate the central role of their research. The progress of aerial robots continues to amaze us.