SBIR Phase I: Extrusion quality inspection with machine learning

Award ID: 2025977

The broader impact of this Small Business Innovation Research (SBIR) Phase I project will be to improve product quality in products generated in continuous extrusion environments. For instance, 2-55% of raw material can be wasted in plastics extrusion. In aggregate at least $500M raw material is lost each year in the US alone, creating additional environmental concerns because this waste plastic is typically not reusable nor recyclable. Manual inspection is problematic for this process at scale. This project will apply intelligent systems to automatically detect and act upon imperfections, improving efficiency and financial performance. The system will initially be applied to plastics extrusion, and later to a wide range of industries including metals, food and pharmaceutical production. This Small Business Innovation Research (SBIR) Phase I project will allow development of novel machine learning and artificial intelligence technologies to automatically detect output of substandard quality in continuous manufacturing environments. The research will generate real-world plastics production data from a range of sensor inputs to train AI models to classify outputs. New approaches in AI/ML will be applied to develop robust, adaptable models to infer error states in product output. Research will also cover the development of technologies to detect failed product output in changing factory conditions, such as fouling of camera lenses or unexpected movement of the hardware. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Program Director: Peter Atherton