SmartDis: Smart disassembly with a knowledge-based automation system

The increasing usage of electronic products is resulting in a growing mountain of scrap electronic equipment on a global level. Disassembling a product into its individual components is considered one of the most important recycling processes in dealing with waste electrical and electronic equipment (WEEE). However, current disassembly processes are primarily performed manually, as the condition of the products provided, which may have damaged or even missing parts, is unpredictable as well as a diversity of product variants in small lot sizes.

The SmartDis research project is focused on developing a knowledge-based, highly autonomous and collaborative robotic system capable of combining non-destructive and destructive disassembly techniques in order to cope with potential process uncertainties. The symbiosis of robotic and human capabilities increases the flexibility of this type of disassembly system. The core components of the system include the decision-making component, the vision system, the digital twin and the execution component, with the latter component being responsible for executing the operations. The decision-making component controls the disassembly process and gradually brings the system from its initial state to the target state (disassembled product). It allocates the disassembly operations to the human or robot, depending on which is most suitable considering time and costs, and it can also create a new disassembly strategy if necessary. The vision system is focused on locating and identifying specific components within the product. Finally, the system status and sensor information is represented by a digital twin in order to support the interaction between the human and the robot, with the human able to use this twin to monitor and support the robotic system, particularly if the robot is unable to complete certain tasks independently. The system's applicability is demonstrated based on a usage case that focuses on the dismantling of antenna amplifiers.

The key to efficient recycling lies in the disassembly of products, which allows certain product parts to be removed and materials to be separated out in order to maximise the recycled resources and minimise the potential for pollution and the quantity of remaining product parts. The SmartDis approach enables faster and easier disassembly of the components within the products to be recycled, thereby ensuring greater gains from the recycling processes. The advantage of this type of approach is that the components can be separated with a high degree of integrity and purity, which enables a higher recovery rate in the recycling processes. Our approach will ultimately enable greater automation in the recycling processes, which may lead to larger parts in waste electrical and electronic equipment being recycled this way in the future.

SmartDis utilises the latest advances in image processing and controls over industrial robots to achieve increased automation, improved data integration, greater flexibility and therefore more efficient disassembly processes. We use an ontology-based product model in this project in order to create a link between the product, the planning processes for the disassembly and the required disassembly equipment.

Our approach uses a captured image of the product to draw conclusions regarding its features and parts, and combines the information extracted with the ontology-based product model. This enables the parts to be recognised and their positions and alignments as well as the path planning to be determined. The robotic system's decision-making component is also able to use the recognised parts to extract the required disassembly information and deduce how a product should be disassembled and how its parts should be removed in succession. These mechanisms are well suited to achieving flexible disassembly processes, but are also beneficial in production systems that need to achieve small batch sizes of down to 1.

We ensure greater flexibility by reducing the resources required to disassembling products with respect to the time and expertise required for configuration and programming. SmartDis will also lay the foundations for intelligent automation to be used in areas that were previously not practical to be automated.

Consortium leader

• Practical Robotics Institute Austria (PRIA)

Additional consortium partners

• Institute of Mechanical Process Engineering and Mechanics, Karlsruhe Institute of Technology
• Automation and Control Institute, Vienna University of Technology
• AUGUSTA Buntmetalle GmbH
• Ing. Eric Dokulil e.U.
• Boxx IT Solutions GmbH
• Reichmann SPS – Service GmbH