2019 2nd IEEE International Conference on Soft Robotics (RoboSoft) | 2019

Nutation Aids Heterogeneous Substrate Exploration in a Robophysical Root

 
 
 
 
 
 
 

Abstract


For exploration of unknown terrestrial environments, it is typically assumed that robots possess sophisticated controllers which can sense important aspects of the terrain (gaps, obstacles, slippery surfaces). However, robust sensing of such conditions is not yet possible in harsh environmental conditions. Biological systems, including plant roots and snakes, are impressive in their ability to use diverse growth and movement strategies to penetrate and explore heterogeneous terrain. Such systems avoid becoming trapped, despite lacking full terrain state information. We are particularly interested in how circumnutation - an endogenous circular pattern exhibited by the tip of a growing root - facilitates penetration and exploration. To discover principles by which robots can gain root-like capabilities, we constructed a planar, pneumatically driven soft-bodied robot, which grows from the tip like a plant root and can bend in 2D space by oscillating the inflation pressure of series pneumatic artificial muscles (sPAMs) arranged on its two sides. We demonstrate that 2D tip oscillation improves the robotic roots ability to penetrate a heterogeneous environment, tested in a lattice of rigid cylinders distributed evenly on a square board. Systematic variation of initial robot positions revealed that the non-oscillating tip strategy led to an increased probability of becoming pinned to obstacles (and preventing growth), while the oscillating tip penetrated the lattice significantly further. The results show that without closed loop control, oscillatory movements of a leading surface of a growing structure enable robust navigation in a heterogeneous environment; closed loop control strategies layered on top of such passive mechanisms could lead to novel strategies for exploratory and search-and-rescue robotics.

Volume None
Pages 172-177
DOI 10.1109/ROBOSOFT.2019.8722717
Language English
Journal 2019 2nd IEEE International Conference on Soft Robotics (RoboSoft)

Full Text