Scenarios for Trust Management in Swarm Robotics

Many terrestrial and aquatic organisms, such as ants, termites or fish, live in communities that resemble larger and more capable organisms thanks to the delicate coordination of the individuals. Humans benefit from mimicking this coordination in many technological sectors, whether it is the coordinated movement of drones, cars, or robot teaming. With the increasing autonomy and intelligence, multi-robot systems and swarm robots specifically have increasing potential to replace humans in dangerous missions, reacting collaboratively to unprecedented conditions. What if, however, one of the robots decides to sabotage such a collaborative mission? How can we support its peers in detecting an untrustworthy member? This paper identifies and classifies the scenarios of swarm-robot collaboration, which is prone to disruption by an untrustworthy member, and links them to a taxonomy of attacks, for which it highlights the importance of the trust aspect between individual robots across the scenarios. The benefits of trust are presented, discussing its capability to prevent attacks and detect malicious individuals in these types of systems.