{"id":126,"date":"2026-02-26T18:03:34","date_gmt":"2026-02-26T18:03:34","guid":{"rendered":"https:\/\/lemursproject.udg.edu\/?p=126"},"modified":"2026-03-18T14:09:02","modified_gmt":"2026-03-18T14:09:02","slug":"progress-on-task-t2-2-ml-based-algorithm-evaluation-for-manipulation","status":"publish","type":"post","link":"https:\/\/lemursproject.udg.edu\/index.php\/2026\/02\/26\/progress-on-task-t2-2-ml-based-algorithm-evaluation-for-manipulation\/","title":{"rendered":"ML-Based Algorithm Evaluation for Manipulation"},"content":{"rendered":"

Within the framework of the LEMURS<\/strong> project, Task T2.2 \u2013 ML-based Algorithm Evaluation for Manipulation<\/strong> has been successfully carried out through two Master\u2019s theses conducted using an industrial manipulator (UR3e).<\/p>\n

This task aims to evaluate the performance of Deep Reinforcement Learning (DRL)<\/strong> and Imitation Learning (IL)<\/strong> algorithms for robotic manipulation, as well as to investigate strategies for transferring learned policies from simulation to real robotic platforms (Sim2Real). In line with the objectives of T2.2, relevant state-of-the-art algorithms were identified, implemented, and experimentally validated on representative industrial manipulation tasks. These include peg-in-hole insertion<\/strong> using DRL approaches, and vision-only assembly<\/strong> using multi-camera feedback through an IL framework based on behavior cloning.<\/p>\n

The work has been developed in the context of the Erasmus Mundus Joint Master Degree in Intelligent Field Robotic Systems (IFRoS)<\/strong> and resulted in the following theses.<\/p>\n

Supervised Learning for Robot Manipulation<\/h3>\n
\n

Author:<\/strong> Tanakrit Lertcompeesin
Supervisors:<\/strong> Narc\u00eds Palomeras and Hayat Rajani<\/p>\n

This thesis explores imitation learning approaches for robotic manipulation, focusing on vision-based policies and industrial assembly tasks.<\/p>\n