Technical Program

Paper Detail

Paper: PS-2A.23
Session: Poster Session 2A
Location: Symphony/Overture
Session Time: Friday, September 7, 17:15 - 19:15
Presentation Time:Friday, September 7, 17:15 - 19:15
Presentation: Poster
Publication: 2018 Conference on Cognitive Computational Neuroscience, 5-8 September 2018, Philadelphia, Pennsylvania
Paper Title: Worminator: A platform to enable bio-inspired (C. elegans) robotics
Manuscript:  Click here to view manuscript
Authors: Raphael Norman-Tenazas, Jordan Matelsky, Kapil Katyal, Erik Johnson, William Gray-Roncal, Johns Hopkins University Applied Physics Laboratory, United States
Abstract: Recent years have seen a renaissance in artificial intelligence (AI) technology and its applications, including robotics. Many of these solutions focus on solving a particular problem in a particular domain or environment. Creating robust and generalizable AI solutions is an area of great interest with applications to many different problem spaces. Biological organisms and biological nervous systems serve as an existence proof that such a generalized intelligence solution is possible. We develop and explore a framework for the simulation of biological networks and extend these simulations to a real-world robotic platform. We focus our initial exploration on a simple, well-defined and highly stereotyped biological neural network (i.e., connectome) derived from the Caenorhabditis elegans nematode. We implement a reference anatomical connectome on a robotic platform, then perturb the network to study the influence of the network parameters on output behavior. Target perturbations can be derived from neuroscience, robotics, or machine learning domains. This platform is useful for exploring the relationship between learning and behavior for biological organisms and robots. To foster further discovery, we share an open source, easy to use framework with visualization and simulation capabilities and provide an interface to a TurtleBot using ROS.