Physics-based and data-driven modeling for biomanufacturing 4.0

Abstract

Bioprinting involves the fabrication of functional tissue constructs using a combination of biomaterials and it has the potential to transform regenerative medicine. However, bioprinting faces several challenges which can be attributed to its high sensitivity to the slightest variation in process parameters, material constituents, and microenvironmental conditions. This research integrates a physics-based model with a memory-based data-driven model to provide predictive capabilities for bioprinting. The hybrid approach uses the long short-term memory (LSTM) network to provide real-time predictions of the bioprinting process parameters as demonstrated by an illustrated case study.