CIRCULAR CIRCUITS: MACHINE LEARNING FOR RECYCLING OF E-WASTE
We are looking for a graduation or internship student in the area of Applied Computer Science or Electrical Engineering for an assignment to be performed within the research project “Circular Circuits”. Here is a link about the overall goal of the project and the various partners within it. The student will work at the Saxion research group Applied Nanotechnology, in close coordination with e-waste companies, on the topic of AI-driven sorting of waste PCBs.
The research group AN is, among other topics, actively involved in the development of machine learning (ML) techniques to allow defect inspection of semiconductor- and MEMS-based devices in the early stage of production. However, we also study if these methodologies can be applied in the characterization of e-waste to facilitate economically and environmentally friendly recycling of the waste.
Specifically, the goal is to investigate whether ML techniques can be of any practical value to delineate regions on a PCB image with similar ‘characteristics’. For example, a preliminary algorithm has been developed that allows to differentiate between the components on a PCB and the PCB itself (Figure 1). The prospective candidate is expected to further develop the technique, as well as look into viable alternatives, e.g. image segmentation (Figure 2). The overall idea is to determine the recycling compatibility and value of a waste-PCB (i.e. “bin it”) by taking an image and then counting and value-weighing each detected / matched component. This will replace costly and slow manual sorting of waste PCBs as is currently done. Adequate and efficient sorting of waste PCBs in different bins allows the metallurgy of the subsequent material recovery step to be optimized, thus costing less energy and less environmental pollution.
Figure 1: Neural Network detection of varied shades of green on a PCB and a discriminated image.
- Understanding the working principles of ML algorithms, specifically with respect to detecting components on waste PCB’s
- Refining our existing ML-model to characterize PCB’s
- Experiments to determine the accuracy of the algorithm on real waste PCB images
- Investigate the viability of using alternative algorithms
Depending on the study direction, background (University or University of Applied Sciences) as well as the nature of the stay (internship or graduation) there are different levels of depth that can be achieved in the above tasks. These are to be discussed with the student’s study advisor before the beginning of the project so as to fit any requirements.
The assignment will be carried out at Saxion within the Applied Nanotechnology Research Group. The graduate will have to attend the regular meetings of the AN-group and is expected to regularly report the progress and discuss issues with the supervisors, as well as with co-workers working on the topic of e-waste.
Are you interested? Send an email to Aleksandar Andreski (email@example.com).