Thesis Work

A dedicated page for master's thesis work and current PhD research direction.

Master's Thesis

Autonomous system design by modeling the problem of passing the front vehicle with deep reinforcement learning method

Öndeki Aracı Geçme Probleminin Derin Pekiştirmeli Öğrenme Yöntemiyle Modellenerek Otonom Sistem Tasarımı

Program
M.Sc. in Computer Engineering
Institution
Erciyes University, Graduate School of Natural and Applied Sciences
Department
Department of Computer Engineering
Supervisor
Asst. Prof. Dr. Fehim Köylü
Year
2023
Status
Completed

My master's thesis focused on modeling the autonomous overtaking problem using deep reinforcement learning and designing an autonomous decision-making system for autonomous vehicles.

Keywords / Methods

  • Deep Reinforcement Learning
  • Autonomous Driving
  • Autonomous Overtaking
  • Curriculum Learning
  • PPO
  • A2C
  • DQN
  • SAC
  • TD3
Current PhD Research

Deep Reinforcement Learning in Contact-Rich Robotic Manipulation Tasks

Program
PhD in Computer Engineering
Institution
Ankara University, Graduate School of Natural and Applied Sciences
Department
Department of Computer Engineering
Supervisor
Prof. Dr. Mehmet Serdar Güzel
Status
Ongoing

My current PhD research focuses on deep reinforcement learning for contact-rich robotic manipulation tasks, with an emphasis on robot learning, peg-in-hole/insertion/assembly-style tasks, and real-world or sim-to-real learning settings.

Keywords / Methods

  • Deep Reinforcement Learning
  • Robot Learning
  • Contact-rich Robotic Manipulation
  • Peg-in-hole
  • Insertion
  • Assembly
  • Model-based Reinforcement Learning
  • Model-free Reinforcement Learning
  • Sim-to-real