Research On Stream Reasoning Mechanism And Method For Ship Navigation Safety Assessment Of Ship's Routing System

With the development of the IOT and the popularity of sensors,a large amount of stream data are produceded.Traditional knowledge representation and data reasoning are difficult to handle streaming data.Stream reasoning is a kind of reasoning mode that combines knowledge representation and reasoning,ontology and stream data processing,can express knowledge about different types of stream data and the connections between them as well as process the stream data in real time.In recent years,the Yangtze River have implemented a ship's routing system.With the construction of smart waterways of ship's routing system,the collection of ship navigation stream data is possible.However,the traditional methods for ship navigation safety assessment can't effectively process stream data in real time,and also can't combine with empirical knowledge,which are inadequate for analysis,judgment of the state of safe of ship navigation.The paper proposes a research on stream reasoning mechanism and method for ship's navigation safety assessment of ship's routing system.The results of the research work are as follows:1)Constructing the ontology for ship navigation safety assessment of ship's routing systemThe paper analyzes the characteristics for ship navigaiton of ship's routing system,establishes the initial set of factors and safety level criteria.It reduces the initial set of safety assessment elements by combining with conditional information entropy rough set.Based on safety assessment elements,combing with the knowledge of navigation safety assessment of the ship's routing system,determinng the process for constructing the ontology;and establishing hierarchical relationships between related classes and classes;the object attributes,data attributes,and individual of the ontology are defined;the domain knowledge are converted into the SWRL rule form.Finally,the safety assesment ontology is constructed.2)Designing stream reasoning mechanism for ship navigation safety assessment of ship's routing systemThe stream reasoning mechanism for ship navigation safety assessment of ship's routing system contains stream data representation and safety assessment stream reasoning engine.The stream data representation uses the RDF framework to describe safety assessment data such as AIS data,meteorological data,and hydrological data to adapt to the timeliness of the data processing and to achieve the fusion of different types of safety assessment data.Combing with the safety assessment ontology,the instance ontology of the safety assessment is constructed,thus to realize the fusion of knowledge of ship safety assessment.The safety assessment stream reasoning engine is used to reason the instance ontology of the safety assessment and adopts a combination of description logic and rule reasoning to reason: the use of description logic reasoning on the SR-Tbox in the safety assessment instance ontology,and the use of rule reasoning on SR-Abox in the safety assessment instance ontology.3)Exploring the stream reasoning method for ship navigation safety assessment of ship's routing systemIn order to improve the timeliness and accuracy of reasoning,the method uses Spark Streaming to receive the safety assessment stream data and combines Jena to generate a safety assessment RDF stream.Then,the method uses SPARQL to query the RDF stream and combines with the SR-Abox of the safety assessment instance ontology to create safety assessment instance ontology.Finally,the SR-Tbox's description logic reasoning algorithm is designed,the rule template transformation method of SR-Tbox reasoning results and the rule conversion method of SR-Rules are studied,and the safety assessment stream reasoning algorithm is designed.4)Experimental verification and analysis of the effectiveness of the proposed method of stream reasoningThe timeliness,detection rate and false alarm rate were selected as evaluation indicators.The safety assessment stream reasoning method is compared with the description logic based reasoning method and rule based reasoning method.At the same time,experiments and comparisons with other ship navigation safety assessment methods are conducted;the experimental results show that the method can timely process the stream data and ensure the validity of reasoning results.