空间点目标识别的神经模糊推理系统应用研究

Study of Spatial Point Target Recognition Based on Fuzzy Neural Network

提出了空间点目标在三个不同波段的辐射通量作为特征向量,使用自适应神经模糊推理系统(ANFIS)和动态进化神经模糊推理系统(DENFIS)进行识别.最后与BP神经网络、遗传算法以及遗传-神经算法三种方法进行比较,进行了仿真实验,仿真结果表明ANFIS,特别是DENFIS具有较好的学习能力和泛化能力,较大的提高了目标识别率,能够较好的进行空间点目标的识别.

In this paper ANFIS and DENFIS are presented for spacial point target recognition with irradiances at three different wave band of spatial target as eigenvector for target recognition. Numerical experiment on the two algorithms are conducted. It is demonstrated that ANFIS, especially DENFIS can effectively learn complex temporal sequences in an adaptive way and outperform some existing models.