Lab for human-learning technologies of complex actions
The circular economy, and remanufacturing in particular, enable sustainable and resource efficient production systems. Used products are recovered, disassembled, reprocessed and returned to the production process. Automation of disassembly processes can be achieved using low-cost robots. However, in many cases this is not profitable due to the high level of programming expertise required. An intuitive method of robot programming that can be carried out by non-specialists is “programming by demonstration”.
The aim of the project is to develop a mobile workstation that can be used interactively by students to experience the challenges of robot programming by demonstration. The holistic process includes capturing and analyzing human disassembly actions using machine learning methods, as well as transferring and executing them on robots.
Development of the mobile workstation for capturing and analyzing human disassembly actions and the mobile station for robot execution
Use of various technologies (including RGB-D camera, microphone, eye tracking) for multimodal recording of humans and advanced collaborative robotics
Development of algorithms based on machine learning methods to process the recorded data
Implementation of a demonstration and visualization to experience the challenges of robot programming by demonstration of the disassembly of a used product.
The project has two main target groups:
The interactive experience of robot programming by demonstration is intended to foster an understanding of the processing of sensitive data in accordance with ethical guidelines, while also addressing and alleviating concerns about the recording of individuals. By presenting the mobile workstations at public events, the holistic process of robot programming by demonstration can also be experienced by all interested people.