A Multi-source Information Fusion Method for Error AIS Targets Identification

ABSTRACT

Automatic Identification System (AIS) is ship navigation equipment. Due to the fact that data-link guarantee mechanism has not been fully resolved, AIS sometimes broadcasts error messages and confused the marine administrators and sailors, so it’s important to automatically identify the error messages. In this paper, we proposed an error AIS messages identification framework, which based on prior probability, expert assessment, fuzzy membership degree and Dempster- Shafer’s (DS) evidence combining rule. In the framework, ship speed, course and longitude-latitude position are core indexes to evaluate the confidence coefficient of an AIS message. Field experiment is carried out at Wuhan section of the Yangtze River, by field observation, the vessels’ real behavior and AIS messages are compared to verify the validity of the algorithm. The experiment results show that for normally sailing ships, the algorithm is able to judge the AIS messages accurately. For ships whose sailing condition changes sharply, such as the touring maritime patrol ship, the algorithm’s recognition accuracy is higher than 84.21%.

Keywords:

Target identification, dempster-shafer theory, behavior characteristic modeling, information fusion, automatic identification System.