Format : On Campus
Number of Interns : 1
Duration of the Internship : See dates
Start Date : 20 July
Finish Date : 8 August
Application Deadline : June 5, 2026
Project Supervisor : Barkan Ugurlu
Project Description : This project explores how human competitive decision-making can be modeled and transferred to autonomous mini-sumo robots using Inverse Optimal Control (IOC). Participants will record human-operated mini-sumo matches and collect trajectory, sensor, and control data from a differential-drive robot. The project assumes that human players behave as if they are minimizing an implicit cost function (e.g., avoiding the ring edge, approaching the opponent, aligning direction, minimizing control effort).
Using data-driven methods, these cost functions will be inferred and then embedded into an autonomous controller. The learned strategy will be evaluated against baseline heuristic controllers and human players in competitive matches. An optional extension includes aggregating multiple human strategies to study collective decision-making and comparing individual versus collective performance.
The project integrates robotics, control theory, data analysis, and behavioral modeling in a competitive and experimentally measurable setting.
Research Intern Responsibilities :
The research intern will model the differential-drive kinematics of the mini-sumo robot and help design controlled human-versus-human or human-versus-robot experiments. They will collect, organize, and preprocess trajectory and sensor data recorded during matches. The intern will define a parametric cost function to represent human competitive strategy and estimate its parameters using basic regression or optimization techniques. Based on the learned model, they will implement an autonomous controller and evaluate its performance against baseline heuristic strategies. Finally, the intern will document the methodology and results and present the findings in a final report and demonstration.
Required Skills and Qualifications :
Strong foundation in high school–level mathematics, particularly calculus (derivatives, integrals) and linear algebra fundamentals (if possible). Basic knowledge of physics, especially kinematics and motion, is expected. Experience with Arduino programming, including sensor integration and simple motor control, is required. Familiarity with data collection, structured experimentation, and basic programming is preferred. Analytical thinking, problem-solving ability, and interest in robotics and decision-making systems are essential.
Expected Learning Outcomes :
By the end of the internship, the student will be able to model a differential-drive robot, design and run controlled experiments, collect and analyze sensor data, formulate and estimate simple cost-based decision models, implement an autonomous controller, and evaluate its performance against human strategies. The student will also gain experience in technical documentation and presentation.
Local Events
Private Events
Find Our Wine
