水下目标识别的知识表示及推理  

Knowledge Representation and Reasoning Algorithm in Underwater Target Recognition

作　　者：闫锐兵[1,2] 高颖[1] 王修亮[1] 彭宇馨[1] 

机构地区：[1]西北工业大学水下信息处理与控制重点实验室 [2]中国人民解放军91388部队

出　　处：《计算机仿真》2011年第1期194-197,共4页Computer Simulation

基　　金：国家水下信息处理与控制重点实验室基金(9140C2304100807)

摘　　要：研究水下目标识别问题,信息复杂,具有不确定性。根据水下目标类别繁多并且递归性强,水下目标识别的规则日益变化,规则对目标识别权重不一,针对各因素都影响了水下目标识别的效率,提出设计面向对象的"封装性"知识表示及其推理的方法,大大的增强了系统的可扩充性能,应用模糊逻辑概念模型进行仿真。仿真结果证明,引入"加权模糊"对知识进行模糊推理,通过样本集对其规则的权系数进行调整和确定,能够让规则的推理更加合理化和客观化,达到提高水下识别效率的作用。Underwater target has a great variety and strong recursiveness, the underwater target recognition has an increasingly sophisticated rules, and the target recognition has different weightings. All these factors have influenced the efficiency of the underwater target recognition. This paper designed an object-oriented ＂encapsulation＂ knowl- edge representation, and the extendlbi]ity of this system was greatly enhanced. Simulation results prove that fuzzy reasoning with fuzzy weighted, fixing and adjusting the weight coefficient of sample set rule was more rational and objective.

关 键 词：知识库 加权模糊 面向对象 目标识别 

分 类 号：TH873.7[机械工程—仪器科学与技术]