Using a Robotic Jellyfish Tentacle to Test Theorized Neural Networks of Control
Lead Presenter: Stephen Smith
Additional Presenters: Kazuo Mori PhD, Dan Blustein
Faculty Advisor/Principal Investigator: Joseph Ayers
Method of Presentation: Poster
A robotic plant modeled after the tentacle of a jellyfish was built to test theorized neural networks for control. The isolated tentacles of the jellyfish Sanderia malayensis were studied using neuroethological techniques to better understand the mechanisms of behavior. Two Teflon coated silver wires were placed into the gastric canal of the tentacle via the cut end. The amplified electrical signals from these wires were capture with simultaneous video during responses to both chemical and mechanical stimuli. Repeatable responses of the isolated system lead to the theorized labeled-line and neural network of sensors and connections. The theorized networks were simplified and simulated with sensory information off-line, the membrane potentials of the motor neurons were fed to the actuators of the robotic plant. Testing the theorized networks on the physical system shows that a proximal-propagating wave of contraction was obtainable using a simple network of neurons.