For the first time, we evolved de-novo a self-organized division of labour mechanism. By "de-novo" we mean that we did not include a priori how the global task should be partitioned into smaller subtasks and what the allocation should be over the different subtasks
Our experimental scenario is inspired by a spectacular form of task partitioning found in some leafcutter ants, whereby some ants (“droppers”) cut and drop leaf fragments into a temporary leaf storage cache and others (“collectors”) specialize in collecting and retrieving the fragments back to the nest. In our analogous robotics setup, we used a team of robots simulated in-silico using an embodied swarm robotics simulator and required the robots to collect items and bring them back to the nest in either a flat or sloped environment
This video shows the evolutionary history that lead to the emergence of task specialization in a swarm of robots. From initially random behavior, the robots first evolve generalist foraging. Subsequently, the robots evolve task partitioning, which gets further perfected over the following generations. At the end of the video we show how the controller evolved in a 4 robot team scaled up when tested in a swarm of 20 robots.