RAAK MKB Close Encounters project - Developing guidelines for human robot cooperation. These guidelines are derived from case studies on mobile and fixed robot applications: AGVs and robot arms. Our focus is on modeling and learning the workflow of human robot interaction.
How can people work together with robots? That is the central question that is approached in RAAK MKB Close Encounters project. We address this question in a multidisciplinary way, e.g., together with Fontys Hogescholen Engineering and Mens & Psychologie.
We focus on understanding the required movements of a robot arm in a production process, in which the robot is working together with human workers. The robot is a cobot. Our focus is on modeling the arm of Rethink Sawyer. We are modeling the arm to fit requirements on accuracy of the positioning, userfriendly customization of the cobot program in the workflow. Our next iteration will be about How do you give a robot arm a human way of moving by presenting the movements required in the workflow?
Another line of attention is on human pose recognition and the possibilities of machine learning for human robot interaction. An interaction module for AGVs is develop a proof-of-concept. Further work is done on how to improve reinforcement learning, e.g., on detecting specific people to start a human robot interaction.
01/11/2018 tot 31/10/2020