Ellen Liu, QBioS Thesis Proposal

Quantitative Biosciences Thesis Proposal 
Ellen Liu 
School of Physics

Neuromechanical Control Adaptations in Complex Environments 
Friday, November 3, 2023, at 2:30 pm 
Location: Howey N201/202 
Zoom Link: https://gatech.zoom.us/j/95203611097 
Open to the Community

Advisor: 
Dr. Simon Sponberg (School of Physics)

Committee Members: 
Dr. Daniel Goldman (School of Physics) 
Dr. Young-Hui Chang (School of Biological Sciences) 
Dr. Nicholas Gravish (School of Mechanical and Aerospace Engineering, UCSD)

Abstract: 
Insects are known to be successful locomotors in complex environments with engineers often taking inspiration from biology to improve the performance of robots. The challenge is determining how these successful behaviors arise from highly complex organisms, composed of many different nonlinear and interacting subsystems. Most studies tend to focus on specific subsystems such as visual sensory feedback or individual muscle responses, however it is important to understand how all these subsystems interact to produce the emergent behavior which yields effective locomotion.

This proposal aims to analyze emergent behavior of insects traversing environments of varying complexity and determine performance benefits that may be of use to robotic systems. This will be done by analyzing changes in neuromechanical control regimes.

Firstly, by using the measure of centralization which measures the degree of neural and mechanical coupling of a system, we look at how cockroaches, a successful terrestrial locomotor vary their control architecture with terrain complexity. As shifts to different control architectures could arise from system limitations rather than to enhance effective locomotion strategies, a robotic platform is to be built to assess which of the two explain the variation in control regimes.

Lastly, we look at a different system where we can take further inspiration from biology by taking exact values from the system (thorax stiffness) and apply it to a robotic model. The chosen system is based on flying insects of which extensive work has been done to determine mechanistic models to describe the observed kinematics. In this system we will analyze possible emergent property advantages by varying between an asynchronous, state dependent actuation strategy and synchronous, time dependent actuation strategy.