摘要
借鉴层次分析法相对比较的思想,提出一种确定多传感器可信度的方法.该方法基于传感器提供的信息的相对可信度,对不同目标分别建立多传感器可信度判断矩阵,求得各传感器对该目标的可信度,进而求得各传感器的综合可信度.给出了一个数值算例.根据传感器的可信度,改变径向基函数神经网络的样本空间的大小,分析了训练时间和网络融合性能,验证了方法的有效性和实用性.
A new obtaining approach of the reliabilities in multiple sensors is offered. Referencing the comparison idea of AHP. A reliability judge matrix for each target is constructed based on the relative reliability of information offered by sensors. Then, the reliabilities of all the sensors for each target and their general reliability can be gotten. A numerical example is presented. By Modifing the number of RBF neural network input samples according to the reliabilities of sensors. The efficiency and practicability of this approach have been validated by analyzing the training time and fusion results.
出处
《测试技术学报》
EI
2005年第1期61-64,共4页
Journal of Test and Measurement Technology
基金
国家863计划资助项目(2001AA602021)
