Short-term vessel traffic flow forecasting by using an improved Kalman model

Authors

Wei He, Minjiang University
Cheng Zhong, Wuhan University of Technology
Miguel Sotelo, University of Alcala
Xiumin Chu, Wuhan University of Technology
Xinglong Liu, Minjiang University
Zhixiong Li, China University of Mining and Technology, University of Wollongong

RIS ID

118201

Publication Details

He, W., Zhong, C., Sotelo, M. Angel., Chu, X., Liu, X. & Li, Z. (2019). Short-term vessel traffic flow forecasting by using an improved Kalman model. Cluster Computing, 22 S7907-S7916.

Abstract

Vessel traffic flow forecasting is of significant importance for the water transport safety, especially in the multi-bridge water areas. An improved Kalman model combining regression analysis and Kalman filtering is proposed for short-term vessel traffic flow forecasting between Wuhan Yangtze River Bridge (hereafter WYRB) and the Second Wuhan Yangtze River Bridge (hereafter SWYRB). Given the vessel traffic flow of WYRB is positively correlated with that of SWYRB, its regression coefficient is obtained as well as the regression predictions. The predictions are further used to replace the state transition equation of Kalman filtering. The prediction results of the improved Kalman model demonstrate better agreements with field observations, and hence, illustrate good capability of the proposed method in the short-term traffic flow forecasting. The discrepancy between the model predictions and field observations is generally attributed to the inherent deficiency of Kalman filtering method and the errors resulted from automatic identification system (AIS) data (e.g. missed AIS data). The proposed method can provide a support for the real-time and accurate basis for the ship traffic planning management.