基于二型模糊系统多源异质信息目标身份类别推理方法

Object Identity Category Inference Method Based on Multi-Source Heterogeneous Information in Type-2 Fuzzy System

针对前件推理规则未知的目标身份识别问题,考虑到目标信息量测的多源异质特点,利用模糊集对目标速度和图像等多维度特征信息进行统一的抽象化描述。借助遗传算法对模糊推理规则进行优化,在确定模糊划分区间的前提下,通过训练得到最优隶属度函数,并建立最优推理规则库。此外,针对因传感器探测精度有限带来的量测不确定性问题,引入区间型样本数据的模糊分类思想,在完成一型模糊系统构建的基础上,提出了一种基于二型模糊推理系统的遗传优化算法。给出了由三角型隶属度函数嵌入构造对应二型模糊集的推导过程,设计了满足点值与区间型数值复合输入的二型模糊推理系统,并通过仿真验证了该推理系统的可行性。

Aiming at the problem of target identification with unknown antecedent reasoning rules, and considering the multi-source and heterogeneous measurement of target information, fuzzy sets are used to describe the multi-dimensional feature information such as speed and image information in a unified abstract way. On this basis, genetic algorithm is used to optimize fuzzy inference rules. On the premise of determining the fuzzy partition interval, the optimal membership degree is obtained by training and the optimal reasoning rule base is established. In addition, aiming at the measurement uncertainty caused by the limited detection accuracy of sensors, and drawing lessons from the fuzzy classification of interval sample data, a genetic optimization algorithm based on type-2 fuzzy inference system is proposed after the construction of type-1 fuzzy inference system. The derivation process of constructing the corresponding type-2 fuzzy set by embedding trapmf membership function is given and a type-2 fuzzy inference system satisfying the compound input of point value and interval value is designed. At last, the feasibility of the inference system is verified by the simulation.