Evolution of Self-organised Path Formation in a Swarm of Robots

We present a set of experiments in which a robotic swarm manages to collectively explore the environment, forming a path to navigate between two target areas, which are too distant to be perceived by an agent at the same time. Robots within the path continuously move back and forth between the two locations, exploiting visual interactions with their neighbours. The global group behaviour is obtained through an evolutionary process and presents emergent properties like robustness, path optimisation and scalability, which recall ants trail formation.

• Type: Journal
• Year: 2010
• Where: ANTSW
• Authors: Valerio Sperati, Vito Trianni, Stefano Nolfi
• Keywords: ANTSW 2010, Artificial Intelligence, Global Group Behaviour, Path Optimisation, Robotic Swarm