Researchers at TU Delft have developed a cutting-edge autonomous navigation system for small drones, drawing inspiration from insect behavior. This 56-gram drone utilizes a combination of visual homing and odometry to navigate effectively. Visual homing relies on visual cues to orient the drone, while odometry measures the distance traveled in a specific direction.
Inspired by how ants navigate using visual landmarks and step-counting, the team created a system that can return home with minimal computational and memory requirements—just 0.65 kilobytes per 100 meters traveled. This innovation makes it possible for tiny drones to perform tasks like warehouse monitoring or gas leak detection without heavy, power-consuming sensors like LiDAR.
The drone tested used a panoramic camera and a microcontroller with 192 kB of memory. It captured snapshots of its surroundings to correct course deviations and navigate precisely. The system’s efficiency lies in high compression of images and minimal processing power, making it well-suited for indoor environments where GPS is ineffective and costly wireless beacons are impractical. This bio-inspired approach, while lacking advanced mapping, provides a feasible solution for autonomous navigation in constrained settings.
