Rats and mice have the ability to move through dark tunnels and crowded burrows by relying on their whiskers to sense and map their surroundings. Now, researchers at Delft University of Technology have shown that a drone weighing under 100 grams can do the same.
In a recent study published in Nature Communications, the team showed that a pair of artificial whiskers allowed a drone to navigate in complete darkness without cameras, LiDAR, or other vision-based sensors, potentially opening the door to navigation in environments filled with dust, smoke, or other visual obstructions.
Too Small for Conventional Sensors
Small drones typically face challenges with sensing, onboard computing, and battery life. The sensors that work well on bigger robots are usually too heavy or consume too much power for drones that weigh less than 100 grams.
Additionally, cameras and LiDAR do not operate reliably in low-visibility conditions, such as smoke-filled buildings or caves. To address these limitations, Salua Hamaza and her team at TU Delft developed a lightweight and low-latency tactile system that enables the drone to navigate without depending on vision.
Inspired By Rodent Whiskers
The Delft University team’s new system attaches a set of artificial whiskers to the front of the drone, inspired by rodent vibrissae. The approach uses biomimicry to adapt a biological sensory system for aerial robotics. As soon as these whiskers contact an object, they provide continuous feedback.
Earlier attempts at contact-based navigation have used larger drones fitted with robotic arms or bumpers. By contrast, Hamaza’s system incorporates real-time tactile sensing into a small flying robot, making touch the primary way of perceiving the environment.
“Here, we aim to equip drones with rich tactile sensing, not for manipulation in the air, but for a novel concept of tactile navigation: using touch to explore and fly through the unknown,” Hamaza said. “But this comes with a challenge: for tactile sensing to work on drones, it needs to be lightweight, low-latency, and low-power. Inspired by nature, we found the answer in whiskers.”
Sensing Contact in Three Dimensions
The drone is equipped with three small pressure sensors at the base of each artificial whisker, which enable it to estimate the three-dimensional position of a contact. The sensors detect pressure changes when a whisker bends against a surface, revealing the location and depth of the contact. Due to this level of detail, the drone is able to avoid obstacles, follow surfaces, and slowly build a tactile map of its environment while in flight.
Filtering Out the Drone’s Own Noise
Distinguishing genuine touch signals from propeller-generated noise is difficult because the airflow causes drift and distorts the tactile data, making it harder to identify real contact events. The researchers developed a real-time processing pipeline that removes the interference and transforms the whisker signals into millimeter-scale depth estimates. The algorithms require only 34 kilobytes of RAM on the drone’s on-board microcontroller.
“We wanted to show that touch does not have to come at the cost of size or computational power,” said Chaoxiang Ye of TU Delft. “Our entire tactile perception pipeline runs onboard using just 34 kilobytes of memory, allowing a tiny drone to sense and respond to its environment in real time.” That pipeline runs entirely on the drone’s own microcontroller, processing tactile data in real time without an external computer.
What the Drone Could Actually Do
During testing, the drone followed both rigid and soft surfaces in complete darkness, using wall-following behavior to explore enclosed areas. As it did so, it gathered information through its whiskers, creating tactile maps of unfamiliar surroundings and locating an exit without cameras or other vision-dependent sensors. The result is a small robot, about the size of a human hand, that navigates an unlit room by touch, much like a rat investigating a hidden cavity in the dark.
In environments obscured by smoke, dust, or other conditions that interfere with cameras and LiDAR, this biomimetic tactile navigation could give small drones another way to sense nearby obstacles and surfaces.
Austin Burgess is a writer and researcher with a background in sales, marketing, and data analytics. He holds an MBA, a Bachelor of Science in Business Administration, and a data analytics certification. His work focuses on breaking scientific developments, with an emphasis on emerging biology, cognitive neuroscience, and archaeological discoveries.