Intention-aware prediction for collaborative collision avoidance in manned-unmanned mixed maritime traffic

• A novel Collaborative Collision Avoidance (CCAS) algorithm named HM-CCAS that enables collision avoidance in Human-Machine and Machine-Machine mixed traffic scenarios. • Adaptation and repurposing of Dynamic Bayesian Network based intention model for the task of predicting the human input in HM-CCAS. This includes the use of future trajectories from the CCAS, rather than using merely the current navigation states. • Verification and validation of the proposed algorithm through a comprehensive simulation study. Communication between vessels is critical for successful collision avoidance. In conventional marine navigation, where all vessels have human crew on board, such a task is trivial. However, with the rise of interest in autonomous vessels, we could expect marine traffic scenarios where both manned and unmanned vessels co-exist for a long period of time until the entire maritime industry fully transitions to autonomous vessels. Implementing a collaborative collision avoidance algorithm in an unmanned vessel in such an environment could be challenging due to the lack of modern computational equipment on board a traditional manned vessel. In this paper, we present an approach that allows collaborative collision avoidance between manned and unmanned vessels. We extend an existing collaboration framework designed for unmanned vessels utilizing a Dynamic Bayesian Network (DBN) to model the response behavior of the manned vessel, enabling manned and unmanned vessel collaboration. A simulation study is used to verify the applicability of the proposed approach and to evaluate its performance.