Interactive & Emergent Autonomy Lab

Giorgos ('Yorgos') Mamakoukas

Research

My research focuses on algorithmic development for underwater robotics, focusing on questions related to efficient algorithms for system identification, nonlinear control, and environmental exploration. More specifically, I have developed methods for controllability-based nonlinear control with guarantees for convergence and obstacle avoidance and techniques for linear embedding of nonlinear dynamics with explicit error bounds on model accuracy for the purposes of data-driven control. I enjoy investigating the properties and relations between different algorithms, such as single-action policies and traditional gradient descent, or feedback linearization and Koopman-operator based control. I am currently working on physics-based identification and control of dynamics.

Education

Certificate, Management for Scientists and Engineerings, Kellogg School of Management, 2018
M.S. Mechanical Engineering, Northwestern University, 2017
B.A. Physics, Grinnell College, 2014

Awards and Honors

Brady Scholar Fellowship, Northwestern University, 2016-2019 H. George Apostle Prize in Physics, Grinnell College, 2014 George White Academic Scholarship, Grinnell College, 2010-2014

Leadership

Introductory Fluid Mechanics, Spring 2019, Grader
Machine Dynamics, Fall 2017, Grader
Machine Dynamics, Fall 2016, Teaching Assistant

Teaching

Introductory Fluid Mechanics, Spring 2019, Northwestern University, Grader
Machine Dynamics, Fall 2017, Northwestern University, Grader
Machine Dynamics, Fall 2016, Northwestern University, Teaching Assistant
Modern Physics, Fall 2014, Grinnell College, Teaching Assistant
Mechanics, Spring 2013, Grinnell College, Teaching Assistant
Modern Greek, Fall 2011 & Spring 2012, Grinnell College, Instructor

Interests

Machine Learning & Data Science, Soccer, House Music, Video Editing, Photography, Road Trips

Projects

Active Learning and Data-Driven Control
Control for Nonlinear and Hybrid Systems
Reactive Learning in Underwater Exploration


Publications

Learning Memory-Efficient Stable Linear Dynamical Systems for Prediction and Control
G. Mamakoukas, O. Xherija, and T. D. Murphey
Neural Information Processing Systems (NeurIPS), 2020.

Feedback synthesis for underactuated systems using sequential second-order needle variations
G. Mamakoukas, M. Maciver, and T. D. Murphey
International Journal of Robotics Research, vol. 37, no. 13-14, pp. 1826–1853, 2019. PDF

Local Koopman operators for data-driven control of robotic systems
G. Mamakoukas, M. Castano, X. Tan, and T. D. Murphey
Robotics: Science and Systems Proceedings, 2019. PDF

Adaptive single action control policies for linearly parameterized systems
O. Ennasr, G. Mamakoukas, M. Castano, D. Coleman, T. D. Murphey, and X. Tan
ASME Dynamic Systems and Control Conference (DSCC), 2019.

Superlinear convergence using controls based on second-order needle variations
G. Mamakoukas, M. MacIver, and T. D. Murphey
IEEE Int. Conf. on Decision and Control (CDC), pp. 4301–4308, 2018. PDF

Ergodic exploration for adaptive sampling of water columns using gliding robotic fish
O. Ennasr, G. Mamakoukas, T. D. Murphey, and X. Tan
ASME Dynamic Systems and Control Conference (DSCC), p. V003T32A016, 2018. PDF

Feedback synthesis for controllable underactuated systems using sequential second order actions
G. Mamakoukas, M. MacIver, and T. Murphey
Robotics: Science and Systems Proceedings, 2017. PDF, Video

Controlling simulated underactuated underwater vehicles with added mass and velocity drift using sequential action control
G. Mamakoukas, M. MacIver, and T. D. Murphey
American Controls Conf. (ACC), pp. 4500 – 4506, 2016. PDF