Development of a Low-Cost Vision System for CAD-Based Guidance of Industrial Robots

Visual Servoing (VS) is an approach that integrates vision systems, i.e., cameras, with robotic arms to provide flexibility to operations such as pick and place or machining that require good positioning accuracy. Also, VS can be used in flexible robotic manufacturing to automatically identify and change the required tool for a specific operation. This paper proposes a vision system based on low-cost depth cameras with an eye-in-hand configuration mounted on a KUKA KR210 R2700 prime with 6 degrees of freedom. The alignment of the real target to the nominal one is based on the iterative closest point algorithm. Two case studies have been developed to test the reliability of the system. First, pick and place tasks of a small block in different operating conditions were conducted. After, the vision system was used to guide the robot in grasping a tool changer. In both cases, the alignment tasks were executed, ensuring sufficient precision.