Prediction Of Traffic Flow In Port Waters Based On Cellular Automaton And Probability Model

With the continuous strengthening of trade cooperation between China and other countries in the world,the density of ships in the port area continues to increase,which is accompanied by an increase in the complexity of traffic conditions in the port area and an increase in the risk of ship navigation.The scientific and accurate ship traffic flow prediction model can provide data support and theoretical basis for maritime agencies and port shipping departments to carry out port infrastructure construction and port traffic flow organization.At the same time,it is the great significance to reduce the occurrence of marine traffic accidents in port waters.The traffic flow prediction model of port water area,which is based on cellular automaton and probability model,uses cellular automata theory and probability models to predict traffic flow for the entire port area.The model uses the historical AIS data mining of the port to combine the actual navigation conditions of the port,and divides the entire port area into two areas: inside and outside the waterway.According to the characteristics of different traffic flows in the two regions,the above two traffic flow prediction models are used to predict the traffic flow in the port waters,and the two parts are integrated into a unified traffic flow prediction platform to interact the two prediction systems.The main work of this article is as follows:(1)The port area is divided into two areas inside and outside the channel by processing historical data combined with actual navigation conditions.The historical data processing is mainly to perform track clustering.Through track clustering combined with actual port area and channel division,some track classes on the chart are divided into channel regions.Channels are drawn on the cluster,and the area outside the chart is uniformly divided into areas outside the channel.(2)The application of cellular automata theory to traffic flow prediction modeling for the area within the channel includes the following steps: Cell behavior is established by ship behavior.Cell neighborhood modeling is performed based on ship domain theory.Cell automata model is established in the channel and cell automata update rules are formulated in the channel.According to the established update rules,the acceleration and deceleration rules in the traditional cellular automaton update model are incorporated into the car following rules.The simulation experiment channel is set as a two-way channel with alignment system,so the overtake rule is reset after the overtake rule is added.When a ship outside the channel crosses the channel,the regional update rules are formulated in conjunction with the rules for collision avoidance at sea,and a crossing rule with a predetermined mechanism is proposed.(3)The probabilistic model theory is used to model the area outside the channel.The specific process includes: Analysis of ship behavior based on historical data,establishment of out-of-way probability model,generation of ship movement probability matrix,ship track deduction based on real-time AIS ship information and ship movement probability matrix.The AIS data used to solve the ship’s moving probability matrix here is the data after interpolation,the data interpolation is performed at a fixed time interval and the time interval is the time step of the cellular automaton update.This ensures that the ship’s moving probability matrix obtained from the data is the probability matrix at the time step.When the vessel crosses the channel,it will cross or wait for crossing according to the feedback signal of the crossing rules in the channel.Through the above studies,the characteristics of the cellular automaton model,strong regularity and stable prediction accuracy,and the probability model,high prediction accuracy and suitable for short-term prediction,are matched to the traffic flow prediction of different characteristics inside and outside the channel.Combining the two prediction models,the prediction accuracy of traffic flow in port waters has been improved to a certain extent,and it provide a new idea for the study of port water traffic flow prediction.