Intelligent ship collision avoidance model integrating human thinking experience

Abstract

With the maturation of intelligent ship collision avoidance technology, the acceptability of this technology has gained attention and the ship collision avoidance model is confronting interpretability and trustworthiness difficulties. Therefore, this paper integrates human thinking experience into actual collision avoidance decision-making to propose an intelligent collision avoidance model that is more in line with the operator ‘s understanding, thus improving trustworthiness and autonomy. The general architecture of collision avoidance model is divided into two modules: a basic model and an analytical model, by adopting from human thinking patterns of fast and slow. The basic model considers the COLREGS and navigation experience to solve the encounter situation of two ships. The analytical model addresses the collision avoidance issues in complex scenarios with multi-ships, by introducing the repulsive force and the driving force parts of the social force model, which are used to implement course mapping adjustment and gentle course fluctuation respectively. Additionally, a field theory-based effect evaluation model is designed to blend human experience with algorithmic benefits for making the optimal decision in complex scenarios. The results show that when the basic model and the analytical model work together, this architecture performs effectively in a variety of scenarios. It is worth mentioning that the collision avoidance model can potentially be developed to advance collision-free and human-like process trajectory navigation in multi-ship situations.