Mining Channel Water Depth Information From IoT-Based Big Automated Identification System Data for Safe Waterway Navigation

Internet of Things technology has been widely used in water traffic research. Many critical waterways in the world are becoming more crowded due to many factors in waterway environments, such as invisibility, variability, and uncertainty. Accurate water depth information is necessary to improve navigation safety. Water depth information of electronic charts cannot be updated in a timely way, while the actual water depth is unpredictable, and this factor threatens the safety of vessels in waterway environments. Based on the shore-based network, ship navigation data and other big data can be integrated to vessels navigation environments in real time. In this paper, we present a new scheme to quickly and accurately construct a vessel safety navigation depth reference map, which contains appropriate channel water depth information. This effective scheme is based on automated identification system (AIS) data and increases the travel safety through crowded waterways. AIS data include rich maritime traffic information. Both the static and the dynamic information about vessels through waterways can be extracted and processed from big real-time AIS data. Based on extensive actual experiments, we apply data mining techniques to extract the waterway depth information both draft-depth and vessel trajectories based on AIS data. The data are collected from vessels in both locations: 1) the Nantong port, in Jiangsu Province, China and 2) Meizhou Bay waterway, in Fujian Province, China. The Hermite interpolation scheme is used to patch the trajectories of vessels, and the BP neural network model is introduced to predict the maximum vessel draft. Clustering and data fusion methods are employed to construct a vessel safety navigation depth reference map according to the cluster area of vessel trajectories and draft information. The experimental results demonstrate that the vessel safety navigation depth reference map accurately reflects the current water depth profile of channels. This paper can provide accurate and timely channel water-depth information for the vessel navigation and the maritime supervision. The proposed scheme in this paper can also provide reference for trajectory data processing and mining.