摘要
无信标无线传感器网络的传感器节点通常是按照一定的概率,以分组形式部署,为实现其定位和动态节点跟踪,提出了的无信标定位发现策略,在已有的部署之上,建立模型去反映目标位置和监测传感器探测信息之间的内在关系,同时还建立了预测模型来对目标移动方式进行推断。利用贝叶斯理论构造了一个条件概率分布,将以上两种模型相关参数归并起来,并在这个分布上应用最大似然估计(MLE)方法来估测目标的位置。实验结果表明此目标定位策略取得了较好的效果。
A beacon-less location discovery scheme was proposed due to the toUowing observations: sensors are usually deployed in groups and sensors from the same group may land in different locations that follow a probability distribution. With this prior deployment knowledge, a model was built to reflect inherence information between positions of the object and observations provided by alert sensors. Also a prediction modal was set up to offer prior knowledge about moving patterns of the object. Generally speaking, the Bayes rule provides us a suitable way to combine those two kinds of knowledge so as to generate a conditional distribution. By applying the Maximum Likelihood Estimation (MLE) method to this distribution, the object's accurate positions could be estimated. Experimental results prove that the scheme can get better effect.
出处
《计算机应用》
CSCD
北大核心
2007年第8期1835-1838,共4页
journal of Computer Applications
基金
国家自然科学基金资助项目(60241004)
国家973计划资助项目(2003CB314801)
关键词
系统设计
预测模型
贝叶斯理论
最大似然估计
system design
prediction model
Bayes rule
Maximum Likelihood Estimation (MLE)