SWARM SECURITY: EXPLORING THE TRADE-OFFS BETWEEN ROBOT ABILITIES AND SCALE, SWARM PERFORMANCE, AND ROBUSTNESS TO DISRUPTION
Abstract
Swarm robotics allows for the deployment of large numbers of robots in applications ranging from search-and-rescue to environmental monitoring. Swarming relies on simple agents, reacting to their neighbours and local environment, resulting in emergent collective behaviours. The de- centralised nature of swarms should, in theory, make them scalable to large numbers and robust to individual failure. Yet previous work has shown that failure of individual agents, whether due to defective hardware or malignant behaviour, could lead to poor swarm outcome. Disruptions include manipulating an agent s goal, its behaviour, the environment, or communication through the environment [5, 6]. Understanding how to make swarms secure to such disruptions is key to their reliable deployment in reality, and their public acceptance.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 11, 2021
- Source ID
- FA86552017033
Entities
People
- Sabine Hauert
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Bristol