Hierarchical Control and Learning of a Foraging CyberOctopus

Soft Robotics In article number 2300088 by Mattia Gazzola and co-workers, an octopus-inspired control framework coordinates multiple soft arms through a hierarchical approach, where reinforcement learning takes high-level decisions, fast energy shaping controls low-level motor execution, reflexive behavior accomodates for environmental disturbances. So equipped, this octopus succeeds in increasingly challenging foraging scenarios, including arena littered with obstacles 3D space.