This paper proposes an adaptive localization approach for wireless sensor networks based on Gauss-Markov mobility model. In the approach,the perpendicular bisector strategy,the virtual repulsive strategy,and the veloc...This paper proposes an adaptive localization approach for wireless sensor networks based on Gauss-Markov mobility model. In the approach,the perpendicular bisector strategy,the virtual repulsive strategy,and the velocity adjustment strategy are properly combined to enhance localization effciency. The velocity adjustment strategy causes that the mobile anchor node automatically tunes its velocity. The perpendicular bisector strategy locally adjusts trajectory for the mobile anchor node,which ensures that unknown nodes obtain enough non-collinear anchor coordinates as soon as possible. The virtual repulsive strategy impels that the mobile anchor node rapidly leaves the communication range of location-aware nodes or returns to the surveillance region after the mobile anchor node was out of the boundary. Both theoretical analysis and simulation studies show that this approach can increase localization accuracy,consume less energy,and cover more surveillance region during the same period than virtual beacons-energy ratios localization scheme using the Gauss-Markov mobility model.展开更多
该文研究了无监督遥感图像分类问题。文中构造了图像的隐马尔可夫随机场模型(HiddenMarkov Random Fleid,HMRF),并且提出了基于该模型的图像分类算法。该文采用有限高斯混合模型(Finite Gaussian Mixture,FGM)描述图像像素灰度的条件概...该文研究了无监督遥感图像分类问题。文中构造了图像的隐马尔可夫随机场模型(HiddenMarkov Random Fleid,HMRF),并且提出了基于该模型的图像分类算法。该文采用有限高斯混合模型(Finite Gaussian Mixture,FGM)描述图像像素灰度的条件概率分布,使用EM(Expectation-Maximization)算法解决从不完整数据中估计概率模型参数问题。针对遥感图像分布的不均匀特性,该文提出的算法没有采用固定的马尔可夫随机场模型参数,而是在递归分类算法中分级地调整模型参数以适应区域的变化。实验结果表明了该文算法的有效性,分类算法处理精度高于C-Means聚类算法.。展开更多
基金Supported by National Natural Science Foundation of China(60776834, 60870010)
文摘This paper proposes an adaptive localization approach for wireless sensor networks based on Gauss-Markov mobility model. In the approach,the perpendicular bisector strategy,the virtual repulsive strategy,and the velocity adjustment strategy are properly combined to enhance localization effciency. The velocity adjustment strategy causes that the mobile anchor node automatically tunes its velocity. The perpendicular bisector strategy locally adjusts trajectory for the mobile anchor node,which ensures that unknown nodes obtain enough non-collinear anchor coordinates as soon as possible. The virtual repulsive strategy impels that the mobile anchor node rapidly leaves the communication range of location-aware nodes or returns to the surveillance region after the mobile anchor node was out of the boundary. Both theoretical analysis and simulation studies show that this approach can increase localization accuracy,consume less energy,and cover more surveillance region during the same period than virtual beacons-energy ratios localization scheme using the Gauss-Markov mobility model.
文摘该文研究了无监督遥感图像分类问题。文中构造了图像的隐马尔可夫随机场模型(HiddenMarkov Random Fleid,HMRF),并且提出了基于该模型的图像分类算法。该文采用有限高斯混合模型(Finite Gaussian Mixture,FGM)描述图像像素灰度的条件概率分布,使用EM(Expectation-Maximization)算法解决从不完整数据中估计概率模型参数问题。针对遥感图像分布的不均匀特性,该文提出的算法没有采用固定的马尔可夫随机场模型参数,而是在递归分类算法中分级地调整模型参数以适应区域的变化。实验结果表明了该文算法的有效性,分类算法处理精度高于C-Means聚类算法.。
