Design And Implementation Of Knowledge Subsystem Of IoT System

With the rapid development of Internet of Things (IoT), the amount of information generated by IoT is growing in recent years. How to deal with such information effectively has become an important issue for the IoT platform. The existing rule-based processing scheme whose definition of rules is too strict, is very complicated in practice. Therefore, on the basis of existing IoT resource model, this paper proposes a knowledge management subsystem, Knowledge Subsystem, which uses and manages the existing knowledge of the Internet of Things.Knowledge subsystem mainly provides the use and management functions of knowledge. The system works with the IoT system resource model and the PUB/SUB system to provide event processing solutions based on the objectives and reasoning. The main work of this paper includes three aspects.In order to facilitate managers and users to use the IoT knowledge for complex event processing and new application development, this paper designs and implements a unified knowledge system, including the management of the IoT resource model and resource instance, knowledge management of event,mapping management between event and resource, and knowledge management of complex event extraction. Its functions include graph-based knowledge modeling, knowledge storage, graph-based knowledge querying, object-oriented knowledge management, and implementation of the association between the user visualization knowledge and the ontology knowledge of the underlying storage by adding an intermediate index layer.The system implements the dynamic knowledge management on the basis of the existing IoT resource model, such as knowledge discovery and knowledge extraction, that is, processing the real-time event flow based on the loT resource model, and computing the complex events hidden behind the resource model and events. This paper designs and implements a complicated event processing system, including concurrent scheduling of multiple reasoning units, resource instance work scope discovery, mapping of resource instance and reasoning unit, transformation between event and knowledge, mapping of event and reasoning unit, knowledge transformation of target events, and target event discovery based on SMT reasoning. The system realizes the real-time processing of large-scale event flow and the dynamic knowledge management by means of divide-and-conquer method. The real-time events are introduced into the knowledge subsystem by using the existing communication infrastructure interface of PUB/SUB system in the existing IoT system. The system completes the event format conversion and implements a Browser-Server mode human-computer interface.Finally, we conduct function test and performance test on the system, the test results show that the knowledge management subsystem based on IoT platform can conveniently manage and query the model, process the event information in time, and combine with the user’s requirement quickly reasoning out the results. The realization of this system has some reference value and significance for the knowledge management of IoT.