Sim-to-Real Reinforcement Learning for Dual-Arm Control
A dual-arm robotic framework inspired by human bimanual manipulation is introduced for automated screw fastening in precision assembly tasks. By coordinating two robot arms, the system stabilizes target objects while simultaneously performing fastening operations, offering improved robustness compared to conventional single-arm approaches. The control policy is learned in a simulation environment and successfully transferred to real robotic hardware, demonstrating the applicability of the approach to real-world assembly scenarios.
