Penn Engineering Advances Swarm AI for Real-World Safety

A team of engineering students and researchers from the University of Pennsylvania and two other universities are developing advanced swarm AI systems designed for real-world coordination, competition, and cooperation. This project, spearheaded by Penn Engineering's xLAB, aims to scale AI agent collaboration to ensure safety in dynamic, uncertain environments.

Rahul Mangharam, who leads Penn’s team, emphasized the critical need for agents to understand intent. "Agents must infer what other agents, human or machine, are trying to achieve and adjust accordingly. They must coordinate without centralized control and respond to dynamic, uncertain environments," Mangharam said. This presents significant algorithmic and systems challenges for large-scale deployment.

The initiative focuses on designing distributed algorithms capable of scaling to hundreds or thousands of robots or drones, enabling consistent real-time decisions. By integrating neurosymbolic AI, which combines neural networks with human-encoded knowledge, the agents gain essential context to operate safely within the constraints of physics.