Motion Planning of the Cooperative Robot with Visual Markers

In recent years numerous robotic solutions have been developed to improve production processes. However, most of the robotic arms are applied in large factories where they perform repetitive tasks supporting mass production. The complexity of programming and adapting robots for changing production is still a barrier for the application of robot arms in flexible and small-scale production. In this paper, we present the system which allows programming the robot using visual cameras and a set of visual markers. We develop a method for flexible motion planning for the cooperative robot to quickly define the trajectory of the robot in the 3D space. We present the architecture of the system, calibration method and the performance of the system. We also show how to significantly improve the detection range of the visual markers using Convolutional Neural Networks for object detection.