论文部分内容阅读
Bayes估计算法是静态环境中多传感器数据融合的常用方法,其信息描述为概率分布,为数据融合提供了一种对多源数据优化处理的手段。然而,该算法需要预先给出不同类型传感器观测对象的分布类型和先验似然概率,并要求各个假设事件之间不相容。为此,数据融合中心不得不根据这些不确定性信息进行推理,以达到目标身份识别和属性判决的目的,使得计算复杂性迅速增加。本文详细阐述了Bayes算法的基本思想,结合数据融合过程的需求,从中归纳出该算法存在的局限性,避免这些局限性影响数据融合效果。表明采用Bayes估计算法可以有效地对多源不确定性数据进行融合,并可以适应融合随时间、空间变化的数据需求。
Bayesian estimation algorithm is a common method for multi-sensor data fusion in static environment. The information is described as probability distribution, which provides a means for data fusion to optimize multi-source data processing. However, the algorithm needs to give the distribution types and a-priori likelihoods of different types of sensor observations in advance, and requires incompatibility among the various hypothetical events. For this reason, the data fusion center has to reason on the basis of these uncertain information to achieve the purpose of target identification and attribute judgment, which makes the computational complexity rapidly increase. In this paper, the basic idea of Bayes algorithm is elaborated. Combining with the demand of data fusion process, the limitation of this algorithm is deduced from it to avoid these limitations affecting the data fusion effect. It shows that the Bayesian estimation algorithm can effectively fuse the multi-sourced uncertainty data and adapt to the data requirement of fusion with time and space.