Simple Cellular Automaton Model for Understanding Caterpillar Swarm Locomotion

Sawfly larvae form a multi-layered swarm and move collectively, like a single large caterpillar. It is suggested that they achieve faster collective movement than individual locomotion by utilizing each other as moving walkways or escalators, the behavior called rolling swarm. However, the behavior remains largely unknown, with many aspects, such as aggregation mechanism and motor control, yet to be fully investigated. Understanding this collective mechanism provides valuable insights into animal swarm systems and establishing a design principle for soft swarm robots. This study investigated aggregation mechanisms underlying rolling swarm via mathematical modeling and simulations. We hypothesized that physical interactions between individual caterpillars help form a rolling swarm. Caterpillars riding on another seem to move faster, and the others being ridden on move slowly. This interaction causes a velocity differential between the upper and lower layers and contributes to aggregation. We developed a simple cellular automaton model to verify the hypothesis and succeeded in reproducing a similar behavior that enhanced speed and climbing ability over rough terrain. These results suggest that local information via physical interactions between individuals can form an adaptive swarm behavior.