Scientists Use AI to Decode the Secrets of Flocking Birds and Swarming Robots

training data
Source: Cell | Validation by training the experimental data.

Key Points

  • Scientists have developed an AI that can learn the rules of collective motion.
  • The AI was trained on real-world data, including the flight paths of pigeons.
  • It learned the two basic rules that govern how individuals interact in a group.
  • The technology can be used to program swarms of robots to form specific patterns.

Scientists have developed a new AI system that can learn the hidden rules governing collective motion, a breakthrough that could help us program swarms of drones and gain a deeper understanding of the natural world. The new framework, published in the journal Cell Reports Physical Science, was created by researchers in South Korea and can identify the simple, local interactions that enable thousands of individuals—from birds and fish to robots—to move together in complex, coordinated patterns.

Collective motion is a fascinating phenomenon. Without any central leader, a flock of birds or a school of fish can move as one. Each individual is only reacting to its immediate neighbors, yet a large-scale, orderly pattern emerges. Determining exactly how this works has always been a significant challenge.

To crack the code, the researchers built a physics-informed AI and trained it on real-world data, including the GPS trajectories of actual pigeons. The AI was able to “learn” two basic rules that govern these interactions: a distance-based rule that controls the spacing between individuals, and a velocity-based rule that causes them to align their direction of travel.

By understanding these rules, the AI can not only predict how a group will move but can also be used to control it. The team demonstrated that they could utilize the AI to program simulated swarms to form specific patterns, such as rings or clumps, and even adjust the size of those patterns on demand.

This has huge practical benefits. In robotics, it provides a blueprint for programming swarms of drones or ground robots to form and switch patterns. In biology, it can help scientists test their hypotheses about how real animal flocks work.

EDITORIAL TEAM
EDITORIAL TEAM
Al Mahmud Al Mamun leads the TechGolly editorial team. He served as Editor-in-Chief of a world-leading professional research Magazine. Rasel Hossain is supporting as Managing Editor. Our team is intercorporate with technologists, researchers, and technology writers. We have substantial expertise in Information Technology (IT), Artificial Intelligence (AI), and Embedded Technology.
Read More