UROP Project

Sensor-Based Characterization of Anthropogenic Material Flows with Machine Learning

About half of the human-induced climate impacts and 90% of biodiversity losses are directly related to resource extraction and processing. Trough recycling anthropogenic waste streams, valuable resources can be saved and ecological advantages can be achieved. In modern sorting plants, mechanical as well as sensor-based sorting machines are combined to transform waste streams into valuable preconcentrates for further recycling. To improve the sorting plant performance, i. e. the amount and quality of recovered materials, the application of sensor-technology for in-line material flow monitoring and for advanced sensor-based sorting (SBS) is an important research topic at the Department of Anthropogenic Material Cycles (ANTS).

Task

In this project, an already existed test rig for in-line characterization of post-consumer lightweight packaging waste flow should be developed further. Different sensors (e.g. near-infrared, RGB and 3D laser triangulation sensors) shall be combined to capture detailed particle and material flow characteristics. With different machine learning models, the captured data should be used to describe process relevant material flow parameters like the particle size distribution or material composition.

Requirements

- Research oriented mindset - Strong problem solving skills - Basic knowledge of Python or a similar programming language and/or Machine/Deep Learning is appreciated - Interest in scientific work and journal publications