摘要
为了获得传感器网络中监测目标的准确状态,需要同时考虑多源节点簇信息融合的时间性和空间性.本文提出了一种多源传感器信息的时空两级融合结构.对同一时刻多源节点簇信息,利用D-S证据理论和支持度进行空间融合,对经空间融合后的时间序列,利用模糊积分、支持度和遗传算法进行时间上的信息融合.仿真实验表明,据此形成的分布式多源节点簇信息融合系统具有目标探测能力、抗干扰能力和容错能力.
In order to receive fully the accurate habitus of acquiring information of the measure object in sensor networks, the spatial and temporal characteristics must be taken into account in the organization of the multi-source node cluster information fusion. A spatial-temporal two-layer late-model is depicted. Using D-S evidence theory and support degree fused the input information measured by different sensors from the space domain. And then using fuzzy integral, support degree and genetic algorithm fused the result of the space domain fused in a relative time domain, and decision-making is also clone by a rule of decision-making rules. The distributed multi-source node cluster information fusion system can improve the capability for object detection, resisting the environmental disturbance and fault-tolerance from the simulation result.
出处
《传感技术学报》
CAS
CSCD
北大核心
2006年第6期2727-2731,共5页
Chinese Journal of Sensors and Actuators
基金
武器装备预研基金项目资助(51419050104JB3204)
陕西省自然科学基金资助(2005F22)
空军工程大学工程学院优秀学位论文创新基金项目资助(BC0507)
关键词
传感器网络
时空融合
D-S证据理论
模糊积分
支持度
sensor networks
spatial-temporal fusion
D-S evidence theory
fuzzy integral
support degree
