多传感器信息融合技术综述

A Survey of Multi-sensor Information Fusion Technology

信息融合是信息领域一个前景广阔的研究方向。通过多传感器组网可以获得比单一传感器质量更高的信息。从信息融合的定义诠释了信息融合的内涵,指出信息融合的实质是推理与决策。给出集中式、分布式和混合式三种模型结构,比较它们的优缺点。融合层次对多传感器信息融合性能有重要影响,给出三种融合层次的性能对比结果。分析四种流行的融合处理算法,比较它们的性能。最后从技术发展和应用需求出发,对信息融合的发展作进一步展望。

Information fusion is a wide research branch in information field.By multi-sensor network,we can get high quality information than single sensor.This paper explains internal meaning of information fusion from its definition.It points out that information fusion is just deduction and decision.It gives out three structure models such as central structure,distributed structure and hybrid structure.Their merits and defects are compared.Fusion phases will have important influence on fusion performance.So the paper gives out the comparsion of 3 fusion phases.Four curret fusion processing algorithms are analyzed and their performance are compared.At last,trends for information fusion are expected from technology development and application.