As AI continues to drive the design and development of intelligent autonomous systems, this project is focused on the developing and implementing algorithms for teams of intelligent autonomous systems.
We will be exploring the convergence behavior of different algorithms including the temporal difference learning algorithms for multi-agent reinforcement learning (MARL) under various situations such as time-varying communication, information quantization, the presence of malicious agents, etc. In addition to theoretical work being done to characterize these algorithms, simulations will be performed to show the validity of the mathematical analysis.
Student responsibilities: Assist in developing Python/Matlab code to simulate the behavior of the algorithms being tested under the guidance of a PhD student and faculty principal investigator.
Student Requirements: Be comfortable with coding in Python/MATLAB. Experience with reinforcement learning is ideal.
Outcomes from this work could be applied to the UCF F1-Tenth racing vehicle.
Most of the work can be done remotely (with virtual discussions) and occasional in-person meetings as needed.
Project Dates
Start Date: 1/6/2025 - End Date: 4/25/2025
Students Needed
Type of Project
Individual or team
Student Responsibilities
Student responsibilities: Assist in developing Python/Matlab code to simulate the behavior of the algorithms being tested under the guidance of a PhD student and faculty principal investigator. Student will begin by reading related papers to better understand the problem space, and gradually transition to algorithm design, analysis, implementation and validation.
Time Commitment
10 hour(s)
Student Requirements
Student Requirements: Be comfortable with coding in Python/MATLAB. Experience with reinforcement learning is ideal. There are two focus areas: i) theory, and ii) algorithm implementation (coding). Students should ideally have interest in at least one focus area. Students may be enrolled in the Math, CS, ECE, or Physics programs. Students in related programs are also considered.
Interested in Working With the Following Programs
For EXCEL URE Students Only
Additional Notes
Most of the work can be done remotely (with virtual discussions) and occasional in-person meetings as needed.