A Motion Planner for Growing Reconfigurable Inflated Beam Manipulators in Static Environments

Abstract

Soft growing robots have the potential to be useful for complex manipulation tasks and navigation for inspection or search and rescue. They are designed with plant-like properties, allowing them to evert and steer multiple links and explore cluttered environments. However, this variety of operations results in multiple paths, which is one of the biggest challenges faced by classic pathfinders. In this letter, we propose a motion planner based on A$<^>*$ search specifically designed for soft growing manipulators operating on predetermined static tasks. Furthermore, we implemented a stochastic data structure to reduce the algorithm's complexity as it explores alternative paths. This allows the planner to retrieve optimal solutions over different tasks. We ran demonstrations on a set of three tasks, observing that this stochastic process does not compromise path optimality.