Research Article
Between Simulator and Prototype: Crossover Architecture for Testing and Demonstrating Cyber Physical Systems
@INPROCEEDINGS{10.1007/978-3-642-30422-4_27, author={Tomasz Paczesny and Jarosław Domaszewicz and Przemysław Konstańczuk and Jacek Milewski and Aleksander Pruszkowski}, title={Between Simulator and Prototype: Crossover Architecture for Testing and Demonstrating Cyber Physical Systems}, proceedings={Smart Objects}, proceedings_a={SMART}, year={2012}, month={9}, keywords={pervasive computing cooperating smart objects middleware simulation demonstration techniques}, doi={10.1007/978-3-642-30422-4_27} }- Tomasz Paczesny
Jarosław Domaszewicz
Przemysław Konstańczuk
Jacek Milewski
Aleksander Pruszkowski
Year: 2012
Between Simulator and Prototype: Crossover Architecture for Testing and Demonstrating Cyber Physical Systems
SMART
Springer
DOI: 10.1007/978-3-642-30422-4_27
Abstract
Consider the development of a new middleware targeted at cooperating smart objects. Each smart object should have an embedded node connected to the object’s sensors and actuators. Building a prototype of such a middleware is inherently labor-intensive, especially when it comes to crossing the cyber-physical boundary, i.e., node-to-object interfacing. Also, soon one needs to be able to validate the middleware’s emerging API. Consequently, two separate “products” are usually developed: a programmer-oriented simulator and an actual, node-based prototype. Both are less than perfect for testing and demonstration purposes, and there is hardly any reuse of work invested in producing them. We propose an architecture that enables intermediate, crossover setups combining elements of the simulator and of the prototype. The key idea is system-wide decoupling of the cyber domain from the physical domain, by means of a dedicated entity. The architecture emphasizes incremental formation of testing and demonstration setups, reusability of elements needed to create them, and flexibility in combining those elements. We validate our architecture with a proof-of-concept infrastructure and a number of experimental setups.
