I. Introduction
In Recent years, robots have been playing an increasingly significant role in industrial manufacturing and processing for their high applicability, efficiency, and accuracy of dealing with complex tasks in various scenarios [1]. Robotic pick-and-place assembly is one of the most common tasks applied to industrial systems. For a specific scene, the end-effector (such as gripper) of the robot moves through a preplanned trajectory from one position to another to pick up and place finished or nonfinished assembly products [2]. However, such automatic pick-and-place task still faces numerous challenges in complex scenarios, such as changeable tasks and assembly objects and unstructured environments [3].