Lightweight Development Method For Human-Cyber-Physical Applications

In the computing environment of human-cyber-physical(HCP)integration,the underlying infrastructure consists of three resources : human,machine and object.The services of these resources are heterogeneous,and the number and variety of resources greatly increase the complexity of resource scheduling,making it difficult to efficiently manage these resources.Moreover,the existing software development platform lacks the means of flexible customization and rapid delivery of various HCP resources according to different application requirements and scenarios.These problems restrict the development of HCP applications.Therefore,this thesis studies the lightweight development method for HCP applications.The specific work is as follows.(1)This thesis proposes a software-definition method SD-HCPR(software-definition human-cyber-physical resources)for HCP resources.It provides a resource encapsulation method based on the resource ontology model,so that the HCP resources in the infrastructure can be encapsulated into finer-grained resource components.The resource ontology model is a formal expression of resource concepts and relationships,which realizes interoperability between heterogeneous resources through shared understanding on the resource model.Its core role is to separate domain knowledge and operational knowledge to shield the upper application ’s perception of resource heterogeneity.Most importantly,it provides a human resource model based on agents and tasks to solve the problem of interoperability with other heterogeneous resources when humans passively accept system control and coordination as a service.In addition,the resource model provides a standard semantic interface for resource components,so that resources can be configured and managed in a service manner;(2)This thesis designs a model-driven HCP service orchestration method.It is a software development method combining service-oriented architecture and modeldriven development,and it is also a key technology for low-code development of HCP applications.This method aims to provide a Maa S(model as a service)application delivery method,and a visual service orchestration model BPMN-HCP is designed for this purpose.The model is a meta-model-based BPMN extension method and is consistent with the BPMN model in terms of choreography semantics and visual performance.These characteristics make it not only ecologically compatible with BPMN,but also overcome the shortcomings of BPMN that cannot represent the special concept of HCP service.In addition,this method provides an automatic conversion method from service orchestration model to executable code.Specifically,through the design of compatible conversion to platform-related BPMN,the automatic conversion of domain business process model to general business process model is realized,so as to achieve automatic generation of executable code;(3)Based on the above two methods,a low-code platform for HCP applications is constructed.It adopts two layer architecture,including resource layer and application layer.The resource layer is composed of resource registration and resource management modules,and the application layer is composed of visual modeling terminal,model transformation tool and business process execution engine.The platform takes the intelligent industrial park as the application scenario and the scene-oriented HCP resource linkage as an example to verify the effect of SD-HCPR on simplifying resource scheduling,improving resource utilization and management efficiency,and the availability of BPMN-HCP to realize the rapid delivery of HCP applications.