论文部分内容阅读
能量受限和数据不确定性是传感器网络在现实应用中所面临的重要问题.本文在不确定传感器数据基础上,利用传感器节点不同属性数据之间的关联性建立概率模型,采用粒子滤波技术进行概率推理从而达到节省能量的目的.首先,针对传感器节点不同属性的数据建立概率模型;然后根据传感器节点属性之间的相关性,采用粒子滤波技术通过感知能量代价低的属性值在模型上推理关联的感知能量代价高的属性值,从而达到节约能量的目的;最后,根据传感器节点多个属性值通常符合多元高斯分布,进而采用高斯粒子滤波进行概率推理,提高推理的精度.实验从准确率和运行效率两方面进行评价,结果表明本文所提出的粒子滤波方法处理不确定数据在准确率和运行效率两方面均能达到良好效果.
Energy limitation and data uncertainty are important problems that sensor networks face in practical application.On the basis of uncertain sensor data, a probabilistic model is established by using the correlation between different attribute data of sensor nodes, and the particle filter technology Probabilistic reasoning to achieve the purpose of saving energy.Firstly, a probabilistic model is set up for the data of different attributes of sensor nodes. Then, based on the correlation between sensor node attributes, the particle filter technology is used to inference the model by perceiving the attribute value with low energy cost Finally, according to the multiple attribute values of sensor nodes, it is usually consistent with multivariate Gaussian distribution, then Gaussian particle filter is used to carry out probability reasoning to improve the accuracy of inference.From the accuracy and The results show that the particle filter proposed in this paper can be used to deal with uncertain data in both accuracy and operational efficiency.