Adaptive Hybrid Vision and Force Control for Dual-Arm Robotic Assembly Tasks

In the field of industrial automation, achieving precise force-controlled assembly remains a critical yet challenging task, particularly for traditional position-based robotic systems. This paper introduces an adaptive hybrid visual-force control approach that significantly enhances the precision and reliability of robotic assembly tasks within complex industrial environments. The proposed system integrates advanced visual recognition using the YOLOv8 algorithm for accurate part detection, coupled with force-controlled manipulation via a dual-arm robotic manipulator for dexterous handling. An adaptive-gain based visual servo control is employed during the assembly process to ensure rapid and accurate positioning, while a force exploration strategy mitigates visual positioning errors, guaranteeing high assembly precision. Additionally, the system features a novel implementation of fuzzy impedance control, which stabilizes the assembly force, resulting in smoother and safer operations. Experimental validation demonstrates that the system achieves a positioning accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\pm \mathbf{0.5}\ \mathbf{mm}$</tex> and a force control precision of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\pm \mathbf{1}\ \mathbf{N}$</tex>, confirming the effectiveness in executing complex assembly tasks.