Arslan, R.; Akşit, M. Chapter 7 – Event-based digital-twin model for emergency management, in Emergency Management, in Management and Engineering of Critical Infrastructures, 1st ed.; Tekinerdogan, B.; Akşit, M.; Catal, C.; Hurst, W.; AlSkaif, T., Eds.; Academic Press, pp. 155 - 176, 2023.
30.09.23
Abstract
Current GIS systems are not expressive enough to represent the data that are required in effective emergency management.
For example, GIS models for hospitals, schools, government buildings, factories, roads, bridges, and water, electricity, gas and communication networks are lacking. In general, control systems incorporate models of the actual parameters of the processes that they control. These parameters are compared with the references and control actions are undertaken if there are deviations. Digital twins are suitable in online modeling of the controlled processes since they are kept consistent with the physical realities. As a general architectural style, event-driven approaches are common in realizing such systems.
Current GIS systems are basically dedicated database systems with topology support. They lack the necessary primitives to support event-driven architectural styles. To overcome this limitation, in general, GIS databases are mapped onto “Object Runtime Systems”, in which the necessary events are implemented. Unfortunately, due to the distributed and heterogeneous nature of emergency management systems and lack of standardization, multiple and incompatible object runtime systems are likely to be adopted by the different subsystems of the architecture. This may cause inconsistencies in referring to the shared events as a consequence of simultaneous read-write and write-write operations.
This chapter introduces three novel contributions. Firstly, as an underlying GIS model, CityGML 3.0 is extended with a set of new abstractions to express the necessary data for effective emergency management. Secondly, a model-checking based analysis is carried out with the model of a distributed and heterogeneous emergency management system to find out the possible conditions that cause inconsistencies. Thirdly, an architectural style is proposed to extend CityGML with an event-driven layer supporting atomic publisher-subscriber protocol. To the best of our knowledge, there has not been any work in the literature which extends GIS systems with new data models and event-driven layers for the purpose of emergency management of critical systems.