In the last study published in Science Robotics, scientists from TU Delft attracted inspiration from ants to development Inspired by insects an autonomous navigation strategy for small, light robots. This innovative approach allows robots to return home after long travels, requiring minimal calculations and memory – only 0.65 kilobytes per 100 meters.
Scientists have long admired the extraordinary navigation skills of ants, despite their relatively simple sensory and neural systems. Previous studies, such as a study at the universities of Edinburgh and Sheffield, allowed the development of an artificial neural network that helps robots recognize and remember the routes in complex natural environments by imitating the navigation ability of ants.
In the last study, scientists focused on small robots, weighing from several dozens to several hundred grams, which have great potential for various applications. Their lightweight design ensures safety, even if accidentally with something completely. Their small size allows them to easily maneuver in tight spaces. In addition, if cheap production is established, such works can be used in a large number, quickly covering large areas, such as greenhouses for early detection of pests or diseases in plants.
However, enabling these little autonomous service robots is serious challenges due to their limited resources compared to larger robots. The main obstacle is their ability to navigate independently. While robots can use external infrastructure, such as outdoor GPS satellites or navigation signals of wireless communication, relying on such infrastructure is often undesirable. GPS signals are inaccessible in the room and may be inaccurate in cluttered environments such as urban areas. Installing and maintaining navigation signals can be expensive or impractical, especially in search and rescue scenarios.
To overcome these challenges, scientists turned to nature. Insects, especially ants, work at a distance important for many applications in the real world, while using minimum detection and calculation resources. Insects combine omnipotentia (tracking your own movement) with visually guided behavior based on their low resolution, but a visual system (view memory). This combination inspired researchers to develop new navigation systems.
One of the theories of insect navigation, the “shutter” model, suggests that insects sometimes capture the snapshots of their environment. Later, they compare their current visual perception with these snapshots to navigate at home, correcting every drift that occurs alone. The main view of the scientists was that the snapshots could be arranged much further if the robot traveled between them on the basis of uncommon. Guido de Croon, a professor of drones inspired by BIO and co -author of the study, explained that homing will work as long as the robot ends enough to locate the shutter, i.e. as long as the drift of steametry of the robot will fall into the “catchment area”. It also allows the robot to travel much further, because the robot flies much more slowly during a shutter attack than when flying from one snapshot to another based on oDometry algorithms.
The proposed navigation strategy has been tested on a 56-gram drone “Crazyflie” equipped with a ubiquitous camera. The drone successfully covers the distance of up to 100 meters using only 0.65 kilobytes of memory. All visual processing was supported by a small computer called “microcontroller”, commonly found in inexpensive electronic devices.
According to Guido de Croon, this new navigation strategy inspired by insects is an important step towards the use of small autonomous robots in the real world. Although the functionality of the strategy is more limited than modern navigation methods, it may be enough for many applications. For example, drones can be used to track stocks in warehouses or monitoring crops in greenhouses. They could fly, collect data and return to the base station, storing images an important mission on a small SD card for processing by the server without the need for navigation.
In related research and development, QDat also made significant progress in autonomous navigation systems for drones in DPS environments. Our innovative approach uses advanced AI algorithms, computer vision and on -board sensors to allow drones to navigate and effective action without relying on external GPS signals. This technology is particularly useful in applications in internal environments, both urban and rural areas, and other difficult conditions when the failure of traditional GPS navigation fails.
These progress is a step forward in implementing small autonomous robots and drones, expanding their potential applications and increasing their operational performance in scenarios in the real world.
