Path planning for large ships in inland waterways considering risk assessment of AIS data

Abstract

Safe navigation of ships moving in inland rivers is a primary concern because of the complex waterways. In the existing global path planning algorithms, the ships might be guided into the opposite pathways, increasing the probability of collision, and the risk of collision is insufficiently modeled in the event of overtaking or head-on encounter. In this paper, a risk assessment integrated path planning algorithm based on Automatic Identification System (AIS) data is proposed. The Cumulative Sum (CUSUM) algorithm and the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm are utilized to analyze AIS data and determine the inflection points of the waterway. A waterway inflection points piecewise improved A* algorithm is devised and used to search for the global path that complies with the navigation rules, and then the Dynamic Window Approach (DWA) algorithm is incorporated into the AIS based collision risk assessment so as to account for the maneuvering characteristics and environmental factors in the assessment of collision risk. Comparison of results of different path planning algorithms shows that the proposed method performs better in terms of collision avoidance.