Study On The Approach Of Ensuring AIS Data Availability In Inland Waterway

The inland waterway transportation plays an important role in comprehensive transportation system because of its large carrying capacity,low cost and less land occupation.The intellective and informational technology is always the hot topic in international navigation society.Furthermore,acquisition of the real-time waterway traffic information is the key technology to build an intelligent navigation system.Automatic Identification system(AIS)is usually regarded as the major senser in maritime situational awareness.The availablity of the AIS data gives significant influence on the safety and effectivity of waterway transportation.However,there are several problems about AIS platform and AIS data,such as low reliability of AIS communication link,frequent appearance of error data and lost data.These problems severely interfere with the maritime situation awareness and vessel collision avoidance,in which decisions are made basing on AIS data.In this paper,we focus our research on the availability of AIS data in the inland waterway.Firstly,based on the AIS data collected in the Yangtze River,statistical analysis is introduced to analyze the reliability of AIS data.With the popurse of studying the AIS message loss problem caused by signal attenuation,the AIS signal field strength in mountainous waterway is acquired.Then matched with the simple geography model,the error curve fitting approach is employed to amend the Egli formula.In the last,an optimization model which can accurately evaluate the AIS signal distribution in mountainous waterway is proposed.The experimental results show that the prediction accuracy of the AIS signal field strength is improved to more than 94%,which is 55% higher than original Egli model.The new model can be used as a guidance to deploy the AIS base station in mountainous waterway,and reduce the blind area of AIS base station consequently.Secondly,to address the problem that the AIS dynamic data are unreliable,the error AIS data screening method and assessing approach are developed.First of all,according to the ships' maneuverability in inland waterway,the threshold of the vessels' displacement distance,speed,average speed,acceleration and course changing rate are applied respectively,to design rules for cleaning the error AIS data.Then,refer to the process of artificially recognizing the error AIS data,an identification approach based on the probabilistic inference is proposed.The approach includes five steps which are priori knowledge extraction,evidence modeling,evidence reliability evaluation,evidence combination,and weight optimization,respectively.The velocity,course,and coordinates included in AIS data were artificially identified,and transformed into belief of evidence between 0 and 1 based on likelihood modeling.Then,the belief of evidence was combined with the ER Rule.With the validation AIS data samples,the weights coefficient of evidence can be trained based on nonlinear optimization model according to various optimization functions,therefore the flexibility and accuracy of the approach was improved.Field experiments show that the proposed approach is capable of identifying AIS data on a high accuracy close to manual work,and cleaning the error and abnormal AIS data effectively.Thirdly,aim at the problem that a great proportion of the AIS data are lost,an approach is developed to restore the lost AIS data.With a large number of data samples,the accuracy of commonly used interpolation methods is verified.And the best interpolation method is employed to repair the lost AIS data rapidly,which can meet the accuracy requirement of restoring lost data in short distance.Then for the long distance lost AIS data,a data restoration method based on the Least Squares Support Vector Machine(LSSVM)regression model is established by introducing the historical similarity trajectory database.Furthermore,the Particle Swarm Optimization(PSO)algorithm,whose optimization target is minimizing the sum of the mean square error,is adopted to search reasonable model parameters.The experimental results show that the proposed model could effectively restore the lost data for a long distance.Finally,according to the restored AIS data,an approach for evaluating the reliability of AIS base station communication links based on the tempo-spatial distribution of lost AIS data is proposed.And a visual analysis platform of AIS base station is established with GIS.In the platform,the geographic grids are automatically generated and combined with the historical weather information and water level information,the loss rate of AIS data is computed for each grid.Consequently,the reliability of the base station communication links can be evaluated,and provide a guidance for adjusting the deployment of the AIS base station.The research contributes to the two major issues including evaluation of the AIS base station deployment quality and the ensurance of the data availability.And the research provides guidance for optimizing the deployment of the AIS base station,and builds a solid foundation for the application of AIS data at the meantime.