Research On Service-Oriented Model And Service Composition For Cyber-Physical Systems

Cyber-Physical Systems(CPS) are a kind of novel intelligent complex systems with the integration and collaboration of computing, communication and controlling technologies. The tight coordination and coupling between cyber and physical resources in CPS will greatly change the way in which we interact with the physical world as well as the cyber world. As the basis of effective design and development of CPS, system modeling is an important part of the research. Based on service-oriented approach, we can develop a service-oriented model to achieve a task by combining different services provided by various physical entities. However, the novel characteristics of CPS hinder the applying of existing SOA and service composition technologies. So in this paper, an in-depth research is done on the service-oriented model and service composition for CPS.The present research status in domestic and abroad is firstly presented in the thesis, especially the advantages and feasibility of modeling CPS in service-oriented methodology. A service-oriented hierarchical architecture for CPS is proposed, in which CPS is divided into four layers:node layer, network layer, resource layer and service layer. On this basis, a CPS service model OWL-CPS is developed in OWL by the inspiration of traditional Web service model OWL-S. The model consists of three upper ontology:OWL-CPSpe, OWL-CPSs and OWL-CPSt, which are able to model physical entities, their services and tasks respectively. AI planning technique, which is used to achieve the goal of service composition, is also involved. The shift mechanism between PDDL and CPS service model is established to convert the composition problem into a planning problem.Eventually, an application scenario of national park is created to illustrate the model. In this scenario, it is assumed that a cyber-physical system is deployed to perform the fire fighting task automatically during the dry season. This CPS is described by the CPS service model OWL-CPS and the service composition process is simulated by software. The outcome shows that this CPS service model and AI planning-based service composition mechanism successfully solved the challenge of PE, service and task modeling as well as the automatic service composition in a novel way.