Cat-Like Whiskers Help Small Drones Fly in the Dark

Researchers Look to Nocturnal Animals to Help Rescue Drones Overcome a Key Challenge: Flying in Darkness

Ultra-compact UAVs weighing less than a stick of butter do not have the computing capacity or power of larger models, and this limits the power of their sensors. This means that if a building fills with smoke or a disaster occurs at night, the drone's cameras and lidar will fail. Even a clear glass can be an obstacle that the drone may not detect.

Now, a research team at Delft University of Technology in the Netherlands - inspired by the whiskers of nocturnal animals - has developed a lightweight, tactile sensor for drones to navigate even when they can't "see". Some nocturnal animals, such as mice and to some extent cats, can explore their surroundings using only the sensations of their surroundings on their whiskers, enabling them to move without sight. The research team wondered if drones could do the same.

In the world of drone development, touch navigation is a new concept that comes with its own challenges. "For the sense of touch to work on drones, it needs to be lightweight, low-latency, and low-power," Delft University of Technology professor Salva Hamaza explained in a press release. Fortunately, nature has already made some of these innovations.

The "whiskers" that Hamaza and his team built for the drone are made of thin nickel-titanium wires, each 0.4 mm in diameter, about the thickness of four sheets of paper. However, the antennas were taller than the drone itself. To the palm-sized drone, the researchers attached two whiskers, each nearly 8 inches long, with three tiny pressure sensors at the base. These can assess the bending of the wires, which allows the drone to estimate its distance and position in relation to nearby obstacles. All calculations necessary for this navigation are processed by a microcontroller on the drone without relying on an external computer.

It may not sound very complicated, but even determining the angle of installation of the wires was a challenge. Animals that move by means of whiskers usually do not fly. Having a long overhang, even if it is very light and delicate, can make the drone unstable. While moving along the wall, the researchers noticed that the resistance of the whisker could cause the drone to spin and fall.

The researchers compared a total of 72 flights and tested six different angles at three different speeds. When the whiskers were installed at 30°, 45° and 60°, the drone flew without significant incident. Of these, an angle of 45 degrees was adopted because it resulted in the least deviation in the direction of the UAV. They also considered the fact that the 60° angle reduces the forward sensing range.

The researchers also encountered several obstacles in processing the information received by the whiskers. One was distinguishing between actual contact signals and disturbances caused by the drone itself, such as propeller vibrations. Even when filtering out this noise, about 38% of drone detections were false positives.

It appeared to be drift, a phenomenon in which sensor readings change gradually over time. Therefore, the research team developed a correction algorithm that continuously estimates drift trends over time—based on sensor readings when the drone is not making contact with anything—during flight. When combined with the filter, false positives were reportedly reduced to 0%.

These experiments were mainly carried out in an environment with three transparent walls arranged in a row. Whenever the tactile sensor touched the surface of the wall, the drone flew parallel to the wall and maintained a distance of 60-100 mm from the surface until it reached the edge. The drone was even able to move when the wall was at an angle to it, something a drone with laser sensors can't do.

However, in a real-world situation—such as the debris of a disaster—obstacles are not ordered in this way. Even in your typical room, a wall-following drone can easily get stuck in a corner or ignore side obstacles. Therefore, the researchers coded a heuristic method that prioritizes movement in unknown regions by probabilistic estimation of the shape of walls. Using this method, the drone successfully found a 2.5-foot-wide exit in a small glass-walled room in 90 percent of tests.

And given that the touch system only uses a small portion of the drone's 192 kilobytes, there is room for improvement, both literally and figuratively. "We wanted to show that touch doesn't come at the cost of size or computing power," added Chao Shan Ye, a PhD student at Delft University of Technology, in a press release. Our entire haptic perception pipeline runs on just 34KB of memory, allowing a small drone to sense and respond to its environment in real time.

This story originally appeared on WIRED Japan and has been translated from Japanese.