It has long been recognised that the key to reducing our waste lies in more comprehensive recycling. Based on this insight, complex sorting facilities have been developed which separate plastics, paper, cardboard, glass, cans and so on into their respective categories. Essentially, this is achieved through various screening processes based on the structure and gravimetric properties of the different fractions. These are separated and shredded across various stages. Smart Live, Smart Device, Hazardous Waste, luminous shoe soles, e-cigarettes, water jugs with filter change indicators, wearables, toys – a seemingly endless list of ever smaller, ever shorter-lived, electronic, mobile, battery-powered consumer products. Current sorting plants are at the mercy of this waste – rough handling leads daily to a multitude of fires, some of which are devastating. Although attempts are made in the sorting plants to manually remove hazardous components in an initial stage, one quickly reaches many limits.

• Small hazardous items are easily overlooked.
• Small items slide down the conveyor belts in the sorting systems and become largely obscured
• It is impossible for the operator to be aware of the variety of integrated batteries – particularly when there is an increased incidence of hazardous objects
• High belt speeds in the pre-sorting stage and ‘buried’ hazardous batteries cannot be dealt with manually
• High manual sorting capacities are at odds with cost pressures and automated plant operation

The aim of the project is precisely to overcome these barriers. The solutions to be developed address

• the detection of concealed and hidden hazardous objects (primarily batteries)
• guided and supervised manual recovery
• robot-assisted excavation/removal based on the development of a solution that can be integrated into existing sorting facilities.

The overall objective of the project is to develop a universal sensor-guided robot for flexible recycling tasks that can operate autonomously. Furthermore, the design will take into account additional areas of application as a basis for future developments. Specifically, the project pursues the objectives detailed below:
The starting point is the objective of realising an X-ray imaging chain that can be easily integrated into existing sorting plants and which captures the incoming waste stream via the image sensor. These images are to be evaluated in real time based on an AI model to be developed, and critical components identified. These are also assessed in terms of their location and their hazard potential. Additional sensors provide data to detect objects that are obscuring the view. On this basis, a removal strategy is developed. This involves combining the manual skills of the operator with the capabilities of the robotic system in a collaborative manner. The sub-areas covered include exposure, removal and emergency interaction. As a general requirement, waste streams with speeds of up to 1.6 m/s, average thicknesses of approx. 200 mm and widths of up to 1.6 m are to be taken into account. The necessary functional prototypes can be constructed on a smaller scale, taking scalability into account.

1. Support in selecting an X-ray imaging sensor system suitable for conveyor belts
2. Specification, development and implementation of a solution that identifies the position and type of object within the waste stream and monitors manual intervention
3. Development and implementation of a robotic platform designed for mobile operation on the conveyor system and capable of operating autonomously in a human environment
4. Development and implementation of a gripper system
5. Support with the collection and cataloguing of sensor data for the detection of
   1. hazardous components (batteries)
   2. typical components that cause obstructions
6. Typical components that cause obstructions.
7. Specification and development of an automation solution that plans work schedules based on sensor data, operator input and robot collaboration, reacts situationally to human input and ensures the final removal of hazardous components
8. Integration of the individual modules into a demonstrator that illustrates detection performance and collaborative removal strategies and enables the testing of interfaces and modules in combination

Network Coordinator:

Xicron GmbH
Heyderstraße 8
99099 Erfurt
info(at)xicron.de

Contact person:

Schmalkalden University of Applied Sciences
Blechhammer 9,
98574 Schmalkalden;
Prof. Dr.-Ing. Frank Schrödel
Tel: +49 (0)3683 688-2107