Enabling Dynamic Reconfigurability of Distributed Control Systems using Reference Nets

An original Petri net-based approach to the embedded control systems construction is presented in this paper. It uses high-level Petri Nets for system specification and provides the final system with dynamic reconfigurability features. We strictly follow the model continuity concept (i.e. the main part of the original model of the system remains the same from the conceptual modeling through a simulation to the final implementation). The source model is transformed into a bytecode that is interpreted by a virtual machine in the target implementation. The approach is based on the well-known Reference Nets formalism which provides the necessary means for multi-level structuring of the model, which allows for the dynamic reconfigurability of the system. The presented solution provides the target system with the ability to install, activate, deactivate, and uninstall its subcomponents during the system runtime using a migratable, remotely installable and uninstallable code. The practical usability of the presented approach has been checked on several hardware platforms.