Content-state-driven motility switching in an intestine-inspired soft-bodied robot via decentralised oscillator networks.

Adaptive handling of thick or composition-changing fluids is difficult for conventional pumps. In animals, the intestine addresses this challenge by switching between segmental mixing and peristaltic transport according to the physical state of the contents. We translate this principle into a silicone soft pump composed of four pneumatic chambers, each driven by its own phase oscillator. Two tunable factors govern the collective behaviour: (i) the coupling strength, which attempts to maintain neighbouring oscillators in a travelling-wave relationship, and (ii) the local sensor feedback, which forces each oscillator to correct the deformation error of its own chamber. Numerical bifurcation analysis and time-domain simulations show that when the two strengths are balanced within an intermediate range, the controller first generates an antiphase pattern that homogenises a viscous mixture, and then spontaneously shifts to a quarter-cycle travelling wave that drives the now-fluid contents downstream. We built a physical prototype and experimentally confirmed autonomous mode switching between two glycerol-based fluids of contrasting viscosity. These results demonstrate that a minimal, bioinspired, distributed controller can endow soft devices with adaptive, multifunctional pumping capability, thereby opening new routes to food-processing, biomedical, and chemical-handling systems that operate under uncertain conditions.