Complex Dynamics and Systems: Embodied and Distributed Control, Sensing and Actuation
Abstract
We seek to develop a framework for understanding distributed sensing and control of soft tactile sensors by drawing from the impressive capabilities of insects. Parsing spatiotemporal information sensed in complex environments is a hallmark of biological systems. However, the mechanism behind how biological sensors process vast amounts of information remains poorly understood. A broad class of animals rely on touch sensation for perception. Among these, cockroaches use a pair of soft antennae for tactile exploration and guidance. There are a staggering 270,000 sensors and nerve fibers that transmit information within the antenna [5]. How this rich information is used to guide decision making remains poorly understood. Cockroaches use their antennae to discriminate textures and shapes, identify gaps to traverse, localize features, etc. [6]Ð[12]. Despite advances in the design and manufacturing of insect-scale robots, there has been comparatively little research in the development of soft, distributed robotic multitouch sensors. By combining experiments with theory and robotics, we will learn how cockroaches process information to detect natural tactile features. In Objective 1, we will quantify how antenna mechanics represent tactile features into spatiotemporal patterns. In Objective 2, we will determine how tactile features are translated into neural patterns and tied to behavior in higher brain centers. Methods Our effort will focus on the American cockroach P. americana due to its remarkable ability to navigate in complex environments. In Objective 1, we will first take high-resolution, 3D micro-CT scans of cockroach antennae to determine the tiling and identity of touch sensors and use vibrometry to determine how mechanical stimuli are transmitted along the antenna. Then, using femtosecond laser micromachining, we will use this knowledge to inform the tuning of a robophysical model whose shape and dynamics can be electrically modulated. In Objective 2, we will use high-density multi-electrode arrays to record form the antennal nerve and higher brain centers. Finally, we will develop a new miniature electronic backpack from stimulating touchrelated regions of the brain to link neural circuits to action selection during untethered behavior. Significance By combining experiments in neuromechanics, theory and experimental robotics, the result of this proposal will reveal general biological design principles that may not only inform our understanding of biological sensors in nature but also provide design rules for more efficient design of the next generation of artificial sensors. Further, our work will build a foundation of new scientific knowledge about how insect brains process complex information to guide intelligent decision making, thereby paving the way for the next generation of embodied Artificial Intelligence. The principles revealed here will inspire a new class of terrestrial robots capable of negotiating challenging terrain with antenna-like sensors.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 19, 2023
- Source ID
- W911NF2310039
Entities
People
- Jean-Michel Mongeau
Organizations
- Army Contracting Command
- Pennsylvania State University
- United States Army