Probability Model Estimation of Aberrant Vessel Trajectory Using VTS Data

In this paper, probability model estimations are presented for vessel trajectories passing beneath a sea-crossing long span bridge. Two methods are developed and compared, which are a model based on 3D vessel maneuvering simulations and one using real world Vessel Traffic Service (VTS) data. Both methods are found to be able to represent the human factor and site specific characteristics to vessel location distribution, which is resulted in a biased shape unlike simple probability distribution function typically assumed in design specifications. Nevertheless, there still exits some discrepancies between the results of two methods. Inadequate mass distribution of samples in VTS and 3D simulation is regarded as a reason for the discrepancy. Further studies on VTS data and research on the application for the vessel collision risk assessment will attain safe and reliable bridge design.