An improved consensus data fusion algorithm is presented,and this algorithm induces relative distance and confidence distance on the basis of the variance of the measured error weight algorithm to improve suboptimal fusion result.
针对目前数字滤波算法中存在对先验信息要求苛刻及定义数据间支持度中门限的预先设定问题,在基于测量方差加权算法基础上,引入相对距离和置信距离的思想对其次优融合估计结果进行改进。
In this paper, the confidence distance measure is used to be fusion measure of data fusion.
以置信距离测度作为数据融合的融合度,先计算出置信距离矩阵,然后分析了计算关系矩阵和确定最佳融合数的几种不同方法,通过分析和算例可以看出,应用椭圆曲线表示的支持程度有助于提高融合结果;并讲述了极值原理法、极大似然法、Bayes法和综合支持程度等数据融合方法,算例表明,这几种融合方法都十分有效,在实际应用中应根据具体情况选择不同的融合方法。
An improved consensus data fusion algorithm was proposed in this paper by the definition of a new confidence distance.
通过定义一种新的置信距离,提出了一种改进的一致性数据融合算法。
Based on the concept of confidence distance measure,the confidence distance matrix and relation matrix for multi-odometer data is constructed firstly.
首先基于置信距离测度概念构造了多里程仪测量数据之间的置信距离矩阵和关系矩阵,然后利用有向图方法剔除含有较大误差的或错误的测量数据,最后采用极大似然估计法求解多里程仪测量数据的最优融合值。
This model revises deviation between apriori distributing and factual distributing of target traits through classified confidence distance measure,which improves reliability of the estimate of target attribute.
该模型利用类别置信距离测度修正目标特征的实测分布与先验分布的偏差,提高传感器目标属性估计的可靠性;利用一致性测度和传感器可靠性,实现传感器信息的筛选,有效的处理冲突信息和错误信息。
CopyRight © 2020-2024 优校网[www.youxiaow.com]版权所有 All Rights Reserved. ICP备案号:浙ICP备2024058711号