| With the continuous improvement of distributed computing and network communication technology,Networked Control System(NCS)is developing towards the control intelligent,the distributed in structure,the system liberation and the product integration.In addition to solving the basic problems of information scheduling,transmission delay,communication resource optimization and fault detection,the networked control system also deals with the diversity of the controlled objects,the complexity of the business,the unpredictability of the environment,the coexistence of various communication protocol standards etc.This poses a higher demand for the NCS oriented middleware’s adaptability.As a kind of software reuse technology,middleware can deal with the problems caused by the complexity and heterogeneity of the underlying environment,and provide the development and operation environment for the upper application,and has the characteristics of dynamic reconfiguration,scalability and context sensitivity,can achieve inter-node coordination and system resource sharing.The adaptive middleware for the networked control system separates the control system from the transparent processing of various protocols,the replication of heterogeneous data,the identification and processing of network faults,the deployment of lightweight and other issues,shielding the complexity of the network environment context And heterogeneity,so that the field of network control applications can focus on business modeling,control algorithm design,functional components of the package,and support the network control system node reconstruction,functional reconstruction and protocol stack reconstruction.Based on the methods of software engineering and adaptive architecture,this paper studies the basic theory and key technologies of adaptive middleware for networked control system,including its adaptive component model,micro kernel design,QoS management framework and service dynamic Configuration framework and so on.Specific research results and innovations are:(1)Proposed an adaptive middleware component model(NCOACM)This paper proposes a new class of adaptive component model(NCOACM)for networked control.The concept of distributed resource boundary is proposed and the description of QoS contract is added.;describes the function of component calculation as a process element in Pi calculus,so that the functional description is independent of the implementation of the programming language.The NCOACM component model supports the dynamic reconfiguration of the structure and behavior of the component entities.The environment factors can be explicitly modeled so that the component entities have the ability to adapt to the open,dynamic and changing environment to meet the differentiated business requirements and adaptive of multiple network protocols.(2)Designed the reconfigurable componentized microkernel-NBMMKBy using the idea of inversion control and slice separation designed the middleware microkernel(NBMMK)to realize the life cycle management of NCOACM component instance;designed the component evolution decision algorithm,evolutionary transactional packaging algorithm,evolutionary consistency analysis algorithm and evolutionary integrity analysis algorithm.The semantic connotation of "component container" and "connector" was extended by analyzing the domain characteristics of networked control system to better support the active reconstruction of component.Further,the component management framework was designed as NBMMK kernel service,and other resource management functions were designed as plug-in non-kernel services,making NBMMK lower real-time resource consumption and better scalability and adaptability.(3)Proposed a component-oriented QoS management frameworkThis paper abstracted the environment context factors of NCS,established the domainoriented QoS concept model by using the method of separation of concerns,and designed a mechanism that supports cross-domain QoS policy encapsulation.By extending the OMG interface definition language constructed the component-oriented QoS policy management framework,realized the optimal allocation of network resources,so that the network control application can be more convenient to adaptive to the variety of network protocol stack.Finally,verified the rationality of the QoS policy framework and the validity of the related algorithms in the QoS component resources satisfaction simulation experiments and the component simulation experiments.(4)Designed the service layer architecture and its related algorithmsDesigned the service-oriented middleware business layer architecture,the core affairs,the call logic and process constraints of this layer were abstracted as services,and the networked control application was regarded as the organism composed of several services;designed the service discovery algorithm,the service call algorithm and the service configuration algorithm so that the business layer can adapt to the change of the domain context and can dynamically configure the composition structure and behavior function so that the middleware has the external sense of adaptive ability.Finally,verified the validity of the service architecture and the related algorithms in the service composition simulation based on QoS target optimization.(5)Designed the adaptive middleware prototype systemDesigned an adaptive middleware prototype system with lightweight kernel structure based on NCOACM component model by using OSGi framework.The prototype system was constructed by using reflection technology,including the basic structure of platform abstraction layer,adaptive communication layer,component frame layer and service layer,and supported the across domains mapping of the QoS parameter;designed component lifecycle management,QoS policy management and other core services,other non-core services to plug-in way integrated into the micro-kernel.Finally,evaluated the model and algorithm of prototype system and verified the validity of the theory and method of adaptive middleware proposed in this paper in the simulation experiment of the performance evaluation of the network controller and the simulation experiment of the middleware kernel’s scalability evaluation. |
