Deep Reinforcement Learning for robot manipulation

Robotic manipulation continues to be an active area of research due to its broad range of real-world applications. Among its benchmark tasks, the peg-in hole problem remains particularly challenging, requiring high-precision control under environmental uncertainty. This thesis presents a framework b...

Descripción completa

Detalles Bibliográficos
Autor: Mulia, Vania Katherine
Tipo de recurso: tesis de maestría
Fecha de publicación:2025
País:España
Institución:Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya)
Repositorio:Recercat. Dipósit de la Recerca de Catalunya
OAI Identifier:oai:recercat.cat:10256/28374
Acceso en línea:http://hdl.handle.net/10256/28374
https://hdl.handle.net/10256/28374
Access Level:acceso abierto
Palabra clave:DRL (Deep Reinforcement Learning)
Deep learning (Machine learning)
Aprenentatge profund (Aprenentatge automàtic)
Robots -- Control systems
Sim-to-real transfer
Peg-in-hole task
Robots -- Sistemes de control
Descripción
Sumario:Robotic manipulation continues to be an active area of research due to its broad range of real-world applications. Among its benchmark tasks, the peg-in hole problem remains particularly challenging, requiring high-precision control under environmental uncertainty. This thesis presents a framework based on Deep Reinforcement Learning (DRL) to train a robotic manipulator to autonomously solve the peg-in-hole task. The proposed approach uses curriculum learning to train a single policy capable of handling all phases of the task: approach, contact-based hole search, and insertion. The curriculum is further extended to incorporate observation noise and force penalization, encouraging the emergence of compliant behaviors during contact. Training is conducted in a custom-designed, physics-based simulation environment. Simulation results demonstrate that the learned policy can complete the peg-in-hole task, though it faces difficulties in balancing task success with compliant interaction. To evaluate the potential for real-world deployment, the trained policy is transferred to a physical robot. Tests reveal several sources of sim-to-real discrepancy, particularly in the modeling of contact dynamics. Nonetheless, partial success in real-world trials suggests the viability of sim-to-real transfer for DRL-trained policies. Overall, this work contributes to the understanding of DRL’s capabilities and limitations in solving complex robotic manipulation tasks such as peg-in-hole assembly.