Anomaly detection model of small-scaled ship for maritime autonomous surface ships' operation. (2nd October 2022)

• Record Type: Journal Article
• Authors: Tyasayumranani, Widiastuti; Hwang, Taewoong; Hwang, Taemin; Youn, Ik-Hyun
• Is Part Of: Journal of international maritime safety, environmental affairs, and shipping. Volume 6:Number 4(2022)
• Page Start: 224
• Page End: 235
• Publication Date: 2022-10-02
• Subjects: Anomaly detection -- AIS data -- trajectory mining -- feature extraction -- navigational pattern; Shipping -- Periodicals; Shipping -- Safety measures -- Periodicals; Shipping -- Environmental aspects -- Periodicals; Shipping -- Safety measures; Shipping -- Environmental aspects; Shipping; Electronic journal; Periodicals; Electronic journals; 387.5
• DOI: 10.1080/25725084.2022.2154116
• Languages: English
• ISSNs: 2572-5084

Abstract

ABSTRACT: As the human's role in the operation of maritime autonomous surface ships (MASSs) is concentrated on less manpower, several issues have been raised regarding the capacity of single manpower. This indicates the necessity of developing monitoring technology for abnormal navigational situations to prevent maritime accidents. Since boating under the influence (BUI) of alcohol is one of the major causes of maritime accidents in Korea, this study focused on BUI of alcohol as abnormal navigation to be monitored. The model suggests a methodology for detecting BUI ships based on their trajectory and behavior. The trajectory and behavior-related features are extracted using AIS and geographic information system datasets and clustered to the anomaly and normal navigation patterns. The proposed model can aid the decision-making of humans monitoring the MASS in detecting abnormal ships in the vicinity of MASSs.