Route-aware detection of ship collision avoidance manoeuvres via change point analysis with intention inference

Zhang, B, Rong, H, Li, H and Yang, Z ORCID: 0000-0003-1385-493X (2026) Route-aware detection of ship collision avoidance manoeuvres via change point analysis with intention inference. Transportation Research Part E: Logistics and Transportation Review. p. 105261. ISSN 1366-5545

Preview

Text Route-aware detection of ship collision avoidance manoeuvres via change point analysis with intention inference.pdf - Published Version Available under License Creative Commons Attribution. Download (11MB) | Preview

Abstract

Reliable and human-consistent collision avoidance is essential for Maritime Autonomous Surface Ships (MASS) operating in mixed traffic environments. A key challenge is the extraction of representative critical encounter scenarios from real-world data without relying on predefined risk thresholds or expert-driven rules. This study proposes a data-driven, route-aware framework for detecting ship collision avoidance behaviours and extracting the critical encounter scenarios in which such behaviours occur. The framework comprises three components. First, ship motion patterns are analysed to characterise normal route-following behaviour, enabling collision avoidance behaviour to be distinguished from routine navigation. Second, a unified cost function is formulated to jointly evaluate multiple kinematic features while explicitly incorporating planned route information, enabling separation of route-following course changes from potential avoidance actions. Change Point Detection (CPD) method is then applied to identify behavioural transitions and segment ship trajectories into candidate collision avoidance manoeuvres. Third, an intention-informed encounter identification approach is introduced, where avoidance intent is inferred by comparing observed trajectories with virtual route-following trajectories. This enables confirmation of true collision avoidance behaviour and extraction of the associated critical encounter scenarios. The proposed method is validated using Automatic Identification System (AIS) data from the high-traffic Øresund region, including the Helsingør-Helsingborg ferry corridor. Results on a labelled dataset show strong performance, achieving an F1-score of 0.8203 and 0.7443 in behaviour identification and encounter extraction, respectively. The findings demonstrate that the framework can robustly identify avoidance manoeuvres and capture critical encounters. Moreover, the proposed behaviour identification method is further applied in various navigational environments to verify its generalisation. The proposed critical encounter extraction framework provides empirically grounded insights into MASS development and testing, maritime situational awareness enhancement, and traffic management.

Item Type:	Article
Uncontrolled Keywords:	0102 Applied Mathematics; 0103 Numerical and Computational Mathematics; 1507 Transportation and Freight Services; Logistics & Transportation; 3509 Transportation, logistics and supply chains
Subjects:	Q Science > QA Mathematics > QA76 Computer software T Technology > TA Engineering (General). Civil engineering (General)
Divisions:	Engineering and Built Environment
Publisher:	Elsevier
Date of acceptance:	13 September 2026
Date of first compliant Open Access:	24 September 2026
Date Deposited:	24 Sep 2026 14:57
Last Modified:	24 Sep 2026 14:57
DOI or ID number:	10.1016/j.tre.2026.105261
URI:	https://researchonline.ljmu.ac.uk/id/eprint/29532

View Item