基于自适应模糊系统的船舶碰撞危险评判

Risk Assessment of Ship Collision Based on Adaptive Fuzzy System

本文应用模糊系统方法建立船舶碰撞危险的专家评判系统。该系统用数值化的模糊联想记忆 (FAM)规则表示专家的领域知识并进行模糊推理 ,同时采用神经网络的微分竞争学习算法 (DCL)实现规则的自适应生成。该系统能模仿人类专家迅速作出危险评判 ,并具有学习、更新和完善的能力。

A risk assessment system of ship collision is established by use of a method of adaptive fuzzy system. Numerical fuzzy associative memory (FAM) rules are applied to the expression of human expert knowledge and fuzzy inference. A differential competitive learning approach in the neural network is adopted to adaptively generate FAM rules and learn new knowledge. The system not only makes an assessment immediately as human experts do, but also has the capacity of learning and refreshing.