Facilitating Programming of Vision-Equipped Robots through Robotic Skills and Projection Mapping

摘要

The field of collaborative industrial robots is currently developing fast both in the industry and in the scientific community. Companies such as Rethink Robotics and Universal Robots are redefining the concept of an industrial robot and entire new markets and use cases are becoming relevant for robotic automation. Where industrial robots traditionally are placed behind security fences and programmed to perform simple, repetitive tasks, this next generation of robots will be able to work side-by-side with humans and collaborate on completing common tasks.This thesis investigates methods for fast and intuitive programming and interaction with collaborative, industrial robots. The work is divided into two areas: Vision-enabled robotic skills and projection mapping interfaces. The purpose of robotic skills in general is to allow non-experts in robotics to program robots in an intuitive manner. It is investigated how a skill based architecture can incorporate advanced robot vision capabilities while keeping the robot programming fast and intuitive. Projection mapping, on the other hand, is the technique to project information onto the real world. It is investigated how projection mapping can be applied as part of human-robot interfaces to simplify and improve human-robot interaction in scenarios involving robot programming as well as human-robot cooperation.