A digital twin-enabled environment for end-user development of robot tasks

The paper presents an End-User Development (EUD) environment that supports end users in the definition of pick-and-place tasks for collaborative service robots. The proposed environment targets professional service robotics scenarios, where robots operate in close collaboration with human workers in shared workspaces. The environment features a chat-based interface leveraging OpenAI GPT−4.1 that allows the user to define robot tasks in a guided and intuitive manner. The defined robot task is then displayed in a graphic interface based on Google Blockly, which allows the user to check, modify, or extend the task by direct manipulation of puzzle-like components. Finally, a robot program can be generated, and a simulation of its execution can be obtained through a digital twin of the robot. This final step is crucial for validating the defined robot task and identifying any potential issues that may arise during execution, before the actual deployment of the robot program. A user study was conducted to assess the usability, workload, and overall user experience of the proposed environment, providing empirical evidence of its feasibility and effectiveness. This paper contributes to the research field of end-user robot programming by proposing a human-centered artificial intelligence approach that leverages the power of large language models and digital twins. This approach empowers end users, even without programming or robotics expertise, to conceptualize, validate, and refine robot programs through intuitive interfaces. By placing users at the center of the development process, the approach aims to ensure accessible, efficient, and effective programming experiences tailored to their needs.