Design And Implementation Of Abnormal AIS Real-time Repair Algorithm For Traffic Command System

The control section of the upper reaches of the Yangtze River is also called "one-way channel",and ships can only pass one way in the control section.The management department has a command signal station in the control section of the river.Through the ship traffic command system,the passing ships can be ordered to pass in an orderly manner.The core of the traffic command system is to receive the ship’s AIS signal in the channel,analyze and obtain the dynamic and static information of the ship,and then perform related operations and generate dispatch instructions according to the rules set by the control section.Due to the influence of various factors such as the complexity of the geographical environment and the instability of AIS equipment,there are often a number of anomalies in the AIS data received by the signal station,which seriously affects the accuracy of the command system decision and the safety of ship traffic.Therefore,how to improve the robustness of the ship traffic command system by real-time processing of abnormal AIS data is of great significance to ensure the safety of ship traffic control in the river section.The thesis aims to improve the robustness of the ship’s traffic command system,with analyzing the problems exposed by the system’s long-term operation in the actual production environment as an entry point,and builds an abnormal data processing mechanism from two dimensions: historical AIS data repair and real-time AIS prediction.The main contents of this thesis include:1 After analyzing,synthesizing and summarizing a large amount of abnormal AIS existing in the historical data of the signal station,the abnormal AIS is classified,and an anomaly data screening model based on the ship’s maneuverability is further established.It detects and recognizes data from historical and real-time dimensions and establishes abnormal data tags.2 Based on the characteristics of the controlled waterway,a short-term repair algorithm and a long-term repair algorithm for abnormal AIS historical trajectories are proposed.For short-term abnormal AIS data,the joint algorithm of weighted least squares and segmented cubic Hermite interpolation combined with the special geographical environment of the river section is used to repair.Through the repair of short-term abnormal trajectories,a relatively complete AIS trajectory database can be initially formed.For AIS trajectories with long-term outliers,a three-layer framework that matches similar trajectories is firstly built to provide training data samples for the improved BP neural network based on GA algorithm,and then the trained model can be used to repair long-term Abnormal track point.3 Aiming at the real-time command requirements of ship traffic command system,a real-time repair algorithm based on GA-BP and Kalman simulation is proposed.After real-time detection and identification about abnormal AIS of the traffic command system,Kalman simulation algorithm is used to repair related attribute data according to the type of abnormality.When abnormality is continuous,the GA-BP model based on the historical similar trajectory is called for the recovery trajectory point,and the repair result will be returned to other functional modules.The abnormal repair information and access instruction records are encapsulated and pushed to the data center.The operation and maintenance server parses,stores,and visualizes the data.After integrating the implemented repair algorithm into the traffic command system and conducting preliminary tests on the spot,it has been proved to deal with abnormal AIS and improve the robustness of the ship traffic command system effectively.Through the designed operation and maintenance server,the repair of abnormal AIS and command decision in real time can be observed and evaluated.