Real-Time Vessel Trajectory Data-Based Collison Risk Assessment in Crowded Inland Waterways

With the rapid development of maritime industries, the vessel traffic density has been gradually increased leading to increasing the potential risk of ship collision accidents in crowded inland waterways. It will bring negative effects on human life safety and enterprise economy. Therefore, it is of vital significance to study the risk of ship collision in practical applications. This paper proposes to quantitatively estimate the ship collision risk based on ship domain modeling and real-time vessel trajectory data. In particular, the trajectory data quality is improved using the cubic spline interpolation method. We assume that the ship collision risk is highly related to the cross areas of ship domains between different ships, which are computed via the Monte Carlo probabilistic algorithm. For the sake of better understanding, the kernel density estimation method is adopted to visually generate the ship collision risk in maps. Experimental results have illustrated the effectiveness of the proposed method in crowded inland waterways.