A collaborative research team from the Netherlands and Germany, drawing inspiration from bees, has put forward a 'bee navigation' strategy. This innovative approach aims to assist small drones in successfully finding their way back after venturing far from their base stations.

Traditional drone navigation methods typically hinge on constructing environmental maps. However, this process demands substantial computational resources and memory capacity. In stark contrast, bees rely on visual motion cues and visual memory to navigate back to their hives with remarkable efficiency.

To delve deeper into the mysteries of bees' visual memory, which remains largely uncharted territory, the research team devised a novel strategy. Initially, the drone is trained to fly in close proximity to its base. During this phase, it collects panoramic images of its surroundings. Subsequently, these images are fed into the 'Cyberzoo' neural network, which processes them to accurately estimate the homing direction and distance.

Experimental results have demonstrated the effectiveness of this solution. After achieving successful small-scale indoor homing, the strategy was put to the test in more expansive indoor and outdoor settings. Impressively, during the outdoor trial, the drone soared over 600 meters away from its base and still managed to return home safely, relying solely on a compact 42-kilobyte neural network.