The development of the next generation of intelligent, efficient robots could be influenced by breakthrough tracking technology that provides new insights into how desert ants navigate their complex environments.
The University of Sheffield is part of an international research collaboration. This initiative has created new tracking technology that tracks certain desert ants throughout their entire foraging life, using computer vision, a branch of computer science, to interpret images and videos. The device monitors the movements of an ant from the first time it leaves its nest until it finds a food source and returns to its colony.
According to the new dataset, ants learn very quickly and only remember their way home after one successful trip.
Interestingly, however, their outward trajectory has changed over time, pointing to various methods of exploration and exploitation. The highly sensitive data also revealed an invisible rhythmic activity beneath the surface; this may help explain how ants construct complex search patterns based on the conditions at hand.
The new software is already used by many research groups around the world and is perfect for citizen science initiatives, as it works with a variety of animal species and uses video captured with ordinary cameras. The resulting precise information is essential to understanding how brains allow animals to navigate their complex environments and could serve as inspiration for a new type of bio-inspired robot.
In a study just published in the journal Science Advances, the University's Senior Lecturer in Machine Learning and Robotics, Dr. Michael Mangan demonstrated the new technology and dataset, along with Lars Haalck, Benjamin Risse, Antoine Wystrach and Leo Clement from the Toulouse Center for Integrative Biology, and Barabara Webb from the University of Edinburgh.
The article shows how CATER (Combined Animal Tracking & Environment Reconstruction) uses artificial intelligence and computer vision to track an insect's position in videos captured with ordinary cameras.
The system works in the animal's natural habitat where other systems fail, as it can detect even microscopic objects that are impossible for the human eye to detect. It's also resistant to background clutter, obstacles, and shadows.
Senior Lecturer in Machine Learning and Robotics at the University of Sheffield, Dr. According to Michael Mangan, it took ten years to develop a system that could extract the data, so it can be said to be ten years in the making.
How these insects are able to travel such long distances of up to 50 km in such threatening environments with temperatures above 1 degrees Celsius, Dr. Michael Mangan has always been intrigued.
Distance tracking of desert ants is traditionally done manually with pen and paper, which requires laying a grid of ropes and pegs on the ground and observing the ants' behavior within the grid. Using a Differential Global Positioning System (GPS) is another way to circumvent this, but the equipment is expensive and has limited accuracy.
There are gaps in our understanding of desert ant behavior, as there is no reliable and inexpensive tool for recording precise insect tracks in the field. In particular, how quickly they set visual routes and what techniques they used to potentially facilitate this process.
CATER's new visual tracking system overcomes these challenges by recording high-resolution images of ants in their natural environment and using imaging technology to identify individual ants based on movement alone. The scene is then reconstructed from high-resolution images using an innovative image mosaicing approach. Bridging the gap between field and laboratory research, this innovative method offers new insights into how ants navigate. This type of information is crucial to understanding how animals with pinhead brains navigate so well in their complex environments.
Opteran, the University of Sheffield spin-out company that reverse-engineers insect brains to create highly durable autonomy using affordable sensors and computing, is already turning such findings into commercial solutions.
Inspiration of New Generation Robots From Ants
Dr. According to Mangan, “Desert ants are an ideal inspiration for the next generation of robots, as they travel great distances and harsh environments like other ants without relying on pheromone trails or GPS and 5G like current robots.
“We hope our tool will enable us to build a more comprehensive picture of how insects learn to navigate their habitats, bring in new scientific insights and train engineers on how to build similarly capable artificial systems,” the researchers write.
Source: phys.org/news
Günceleme: 26/04/2023 18:49