Digital twin for scheduling and allocation of real-time cyber-physical systems

Here at the Real-Time and Distributed Systems Group, one of our primary goals is to improve the timing, reliability and resilience of safety-critical systems. This motivates us to apply the latest technology to engineering systems, including those infrastructures to support these systems, e.g. base stations for real-time communication.

In our recent MOCHA project (2019-2023), we utilize a learning-enabled scheduling paradigm with Digital Twin that models the system as it operates and applies a data-driven modelling approach to understand the characteristics of the target system. We then utilise the learnt models and decide when and how it is appropriate to use that model in the decision-making process. 

Our Digital Twin-based scheduling and allocation algorithm has been experimented with by our funder and achieved a 20% improvement, compared to their initial design and will be implemented in the next generation of 5G/6G base stations from our industrial partner.