Mechanism for Constructing the Dynamic Collision Avoidance Knowledge-Base by Machine Learning

Abstract

Machine learning is the key technology for constructing a dynamic collision avoidance knowledge-base and realizing the Personifying Intelligent Decision-making for Vessel Collision Avoidance (short form for PIDVCA). Based on the aim and realization method of PIDVCA, this thesis puts forwards a Meta-Knowledge representation integrated method, which is carried by procedural knowledge, combined with the knowledge of the facts on basis of the database technology and the knowledge of the cause and effect on basis of the formation rule, integrated by the intelligent decision-making chart and tree or rule set, The mechanism for constructing a dynamic collision avoidance knowledge base by the integrated machine learning strategy are discussed by combination of collision avoidance simulation instance. Simulation results show that the dynamic collision avoidance knowledge-base can make the effects of the PIDVCA.