Evolving Functional Self-Assembling in a Swarm of Autonomous Robots

The goal of this study is the design of controllers for robots capable of physically connecting to each other, any time environmental contingencies prevent a single robot to achieve its goal. This phenomenon is referred to as functional self-assembling. Despite its relevance as an adaptive response, functional self-assembling has been rarely investigated within the context of collective robotics. Our task requires the robots to navigate within a rectangular corridor in order to approach light bulbs positioned on the opposite end of the corridor with respect to their starting positions. Aggregation and assembling are required in order to traverse a low temperature area, within which assembled robots navigate more eectiv ely than a group of disconnected agents. The results of our empirical work demonstrate that controllers for a group of homogeneous robots capable of functional self-assembling can be successfully designed by using articial neural networks shaped by evolutionary algorithms.