Safety for human-robot interaction in dynamic environments

Abstract

One of the key issues in new flexible manufacturing set-ups is how the robot system interacts with its environment, specially when people are involved in those interactions. Thus, in order to guarantee safety without restricting robot system autonomy and flexibility, a novel, robust visual application to detect and track people around a robotic system is presented here. First, a new background maintance algorithm is applied in order to adapt to different unexpected changes in the scene. Then, after applying a calibration method, the estimation of the position between the system and the detected objects must be done by using an additional method. This application has been tested and compared with the previous approaches by demonstrating its improvement over them, as it will be shown in the results section.