Artificial force fields for multi-agent simulations of maritime traffic: a case study of Chinese waterway

摘要

Collisions are an important factor in probabilistic risk modeling of waterway infrastructures. Especially in the design of bridges and terminals, the probability of ship collision is important. Researchers are looking for simulations that mimic ship behavior and estimate accident probabilities rather than using probabilistic accident models. AIS (Automatic Identification System) data forms a valuable basis for development of such simulation models. The data is a valuable asset for the validation and calibration of accident simulation models. Hence, a detailed analysis of AIS data is being undertaken as part of current research on ship collision simulations. Based on AIS data analysis, we developed a primary application of multi-agent simulation for ship traffic in the Yangtze River, which provides a realistic representation of the shipping traffic. Visual comparisons between AIS data and simulation data show comparable ship movements. More importantly, many different simulations show that the output of the simulation reproduces the behavior of real ships according to AIS data. Further down the road, ship failures will also be simulated, which results in a probability of ship accidents. Rather than the existing mathematical methods to estimate the risks of maritime accidents, multi-agent simulations provide a realistic ship behavior and all the details of ship movement. The model provides detailed information on how ships go wrong, and forms a useful tool in the design of waterways, design of bridges, and in traffic management.