Bio-inspired sensing and control of soft slithering bodies in 3D environments

ORAL

Abstract

We present a bio-inspired neural architecture for learning closed-loop controls in slithering soft bodies, realized through coupled oscillators and reinforcement learning. We deploy this architecture in a soft snake model and demonstrate how the snake can filter and leverage noisy sensory measurements to navigate realistic 3D environments.

*NSF EFRI #1830881ONR MURI N00014-19-1-2373

–

Presenters

  • Xiaotian Zhang

    • University of Illinois at Urbana-Champai

Authors

  • Xiaotian Zhang

    • University of Illinois at Urbana-Champai

  • Tixian Wang

    • University of Illinois at Urbana-Champaign

  • Andrew Dou

    • University of Illinois at Urbana-Champaign

  • Prashant Mehta

    • University of Illinois at Urbana-Champaign

  • Mattia Gazzola

    • University of Illinois at Urbana-Champaign