Towards Enhanced Vessel Perception in Inland Waterways: Integration of Radar and AIS Data

In inland waterways, characterized by high vessel traffic density and collision risks, there is an increasing demand for improved perception capabilities in vessel navigation. This paper addresses the issues of low tracking and identification accuracy of ship navigation radars and AIS in inland waterways, as well as the challenges of asynchronous data fusion. Firstly, we synchronize the sampling frequencies of radar and AIS target motion data. Subsequently, we employ a fast covariance intersection method to integrate data from multiple sensors, thereby obtaining high-precision target data. Finally, we deploy perception devices along the Wuhan section of the Yangtze River inland waterway, complemented by onboard high-precision GPS, to experimentally validate the accuracy of proposed multi-source fusion method in target recognition and tracking.