Guest Talk – March 29, 2023
Giulia De Masi – Principal Scientist
Collective Intelligence: from biological and social to robotic systems
How a group of individuals reaches a consensus through local interactions and without centralized authority is widely studied in social, biological, political, and information sciences. Strongly opinionated minorities and asocial behaviors can dramatically impact the opinion dynamics of a large population. Here we will explore different scenarios of consensus achievements with asocial behavior, extending them to applications of consensus formation to a system of robots.
The mean-field models and multi-agent simulations first predict the results and are finally confirmed by experiments with swarms of 100 locally interacting robots.
The field of Robotics exploring the collective behaviors of decentralized systems of robots is called Swarm Robotics and is inspired by the collective behavior of social insects (swarms of bees). While the usage of single robots is well established, using groups of robots in a decentralized control setting is an increasingly relevant research topic. Based on the local interactions of a single robot with the environment and other robots, the final aim is to design the emergent behavior of a whole swarm of robots for specific tasks. Compared to a single robot, a swarm is more robust, flexible, efficient, and scalable. Together with simulations, the availability of low-cost microrobots allows the realization and study of a swarm of robots in the laboratory before the real applications on the field.