“Precisely because things went wrong for me, I figured out how nature deals with this problem”

My research into extremely small flying robots is a challenge. There is little room for the large sensors and computing power of our drones, with the smallest weighing just 3 grams. This scale makes it very useful, for example, for “search and rescue” operations in collapsed buildings or for flower pollination. So we try to draw inspiration from nature.

For a while now, I’ve been working with the European Space Agency (ESA) on the routes for these small drones to land. For this I searched the biological literature. Biologists have claimed that during landing, bees benefit from movements in their field of vision, which is called “optical flow”. When you look outside from the train, you can see the trees closer and moving faster than the mountains in the distance. Trees have a greater optical current than mountains.

Since I can easily apply this principle to my own small drones, I quickly arranged a demonstration at a large Esa indoor seminar. Even the great chief of Jesus was present.

During my demonstration, the flying robot delicately flew, and everyone started clapping. But near the ground, the drone began to vibrate up and down, preventing it from landing. The whole audience laughed. It was a complete failure.

After this disappointment, I began to think about this phenomenon. Other researchers have always achieved good results with this technique. So why didn’t I succeed? Perhaps the air was bouncing back when the drone approached the ground or perhaps it was image processing.

Even during the Christmas holidays in France, I kept wondering what the solution was. I have studied landing simulation with small planes. It turns out that the same problem occurred in the simulation. So it wasn’t the wind or the camera.

I once again delved into the literature. This is how I discovered that shaking always happens. Even pilots suffer from this if they fly without all sorts of heavy equipment. Another disappointment, apparently was an unsolvable problem.

Then I thought: Maybe we can use swings. If the drone detects these motions, it knows the Earth is nearby. According to the literature, bees do not see a distance, but they all begin to swing exactly 7 centimeters from the flower. It’s kind of a constant flight correction, don’t I do too much pressure, don’t I get too weak before landing? My discovery was that this swing is necessary for animals to see the distance and land properly.

In the following years I worked a lot with biologists. These biological experiments turned out to be much more difficult than robot research. For example, an Asian wasp ate all of our bees during the experiment. However, we have come up with more ideas about how bees land. This eventually led to it being published in the famous trade magazine The intelligence of nature’s machine.

Other researchers have tried for years to deny this problem, for example by putting long legs under their drone or landing very slowly. But precisely because things went wrong for me, I figured out how nature deals with this problem and my drones can also land better.