Honeybees may hold the secret to the future of drone technology.
Researchers in Europe have developed a new navigation system inspired by the way honeybees find their way home, allowing tiny drones to travel long distances while using very little memory or computing power.
The study, published in the journal Nature, was led by scientists from Delft University of Technology in collaboration with researchers from Wageningen University and Carl von Ossietzky University of Oldenburg.
The team created a navigation system called “Bee-Nav,” which copies how honeybees learn and remember their surroundings.
The goal was to help small robots and drones navigate independently without needing large computer systems or detailed digital maps.
Modern drones can already perform many impressive tasks, from inspecting buildings and monitoring crops to delivering packages.
But navigation remains one of the biggest challenges. Most drones rely on GPS or complicated mapping systems that require powerful processors, large memory storage, and a lot of energy. This makes drones heavier, more expensive, and less efficient.
Honeybees, however, solve the same problem with brains smaller than a grain of rice.
Despite their tiny brains, honeybees can fly far from their hive along winding routes and still return home successfully. Scientists believe they do this using a combination of odometry and visual memory. Odometry is a way of estimating movement by tracking direction and distance, similar to counting steps while walking. But odometry becomes less accurate over time, so bees also memorize what their environment looks like near important places such as their hive.
The researchers designed Bee-Nav around this idea. Just like a young honeybee, the robot first performs a short learning flight close to its home location. During this flight, it captures panoramic images of the surroundings. A small neural network then learns to estimate the direction and distance back home using those images.
Surprisingly, the system required extremely little memory. In some tests, the neural network used only 3.4 kilobytes of memory. Even in larger outdoor tests, the full navigation system required only 42 kilobytes, far less than traditional drone navigation systems.
The researchers tested the drones in both indoor and outdoor environments. In one outdoor experiment at the Dutch drone testing site Unmanned Valley, the drone successfully flew more than 600 meters away and still returned home. Indoor tests were highly successful, while outdoor tests in windy conditions had a success rate of about 70%.
The scientists found that strong wind made navigation more difficult because it tilted the drone and distorted the images it used to recognize its surroundings.
Even so, the researchers believe the technology has great potential. One possible use is greenhouse monitoring, where lightweight drones could safely inspect crops for pests or disease without posing risks to nearby workers.
The study may also help scientists better understand how real honeybees navigate the world. By studying insects more closely, researchers are discovering that nature may already contain elegant solutions to some of technology’s biggest problems.
