This project was carried out under TWI’s Exploratory Programme, the purpose of which is to drive innovation and generate commercially exploitable, new intellectual property (IP) or know-how for TWI. Through the Exploratory Programme, TWI aims to provide funding for concept validation, prototype development, market analysis, and IP development, protection, valuation and exploitation.
PROTEAS leveraged competencies for the global digital twin market by demonstrating a functional proof of concept that empowers the training and deployment of machine learning (ML) models, and establishing an Edge artificial intelligence (AI) framework that enables the efficient utilisation of ML models on digital twins in a distributed fashion. Successful completion of the project resulted in the integration of cutting-edge, cloud-native tools, making ML operations and utilisation of graphics processing units on the Edge for hardware-accelerated training possible. The implementation of PROTEAS has yielded a reduction in the ML lifecycle time by over 20% and an increase in the accuracy of data-driven digital twins by more than 10%, as well as enabled continuous monitoring of digital twin operations, verification of accuracy and retraining if needed.
Additionally through PROTEAS, the project team was able to harness these enhanced capabilities in digital twin and ML to deliver a collaborative project for TWI Hellas, opening up a new domain in the space and data sector, which holds immense potential and offers a new area of opportunity for the future.