RAZER - A HRI for Visual Task-Level Programming and Intuitive Skill Parameterization

Abstract

Maintaining competitiveness and mitigating health issues test caused by unergonomic working conditions are two main reasons for automating production processes. But such automation is expensive, also because test experts are required to program the robots. One approach to lowering these costs is to enable shop-floor workers to program robots by providing task-level programming tools. Task-level programming is an established approach, yet appropriate workflows for experts and shop-floor workers remain to be defined. The objective of this letter is to evaluate RAZER, a framework for robot task-level programming, in which skill programming and parameter interface definitions are integrated. The framework provides workflows for both experts -- creating skills and providing their parameter interfaces -- and for shop-floor workers -- using these skills to create executable robot tasks in an intuitive human-robot interface (HRI). The HRI is a graphical user interface that runs in a browser, and provides access to other man-machine interface, such as programming by demonstration. Two pilot and two user studies proof that RAZER fulfills the demands of both experts and novice users.