一种基于传感器网络的水污染源质心定位算法

摘要

目前基于传感器网络的污染源定位多采用基于扩散模型的解析定位算法.然而,在该类算法中,复杂的数值计算会引入估计误差,并且水污染扩散模型大多是在理想的近似条件下提出的,在污染源定位问题中亦会引入误差.在某些情形下,由于没有解析的定位模型,基于扩散模型的解析定位算法无法采用.为了弥补传感器网络下基于扩散模型的水污染源定位方法的不足,提出了一种不依赖于扩散模型的水污染源质心定位算法.在该算法中,首先确定污染源所在区域,然后计算几何区域的质心,质心位置即为污染源估计位置.为了求解质心定位问题,提出了基于浓度场等位线的求解方法.在实验部分,对比了本文质心算法与基于扩散模型的定位算法以及粗略定位算法的定位结果.实现结果说明了本文算法的有效性.%In the pollution source localization by using sensor networks,analytic localization algorithms depending on diffusion models are always adopted. However,there are errors in the complicated numerical computation,and the diffusion models which are proposed under ideal assumptions bring errors in the localization. In some special cases, there are no explicit diffusion models and thus the localization methods related to diffusion models cannot be ap-plied. To make up for the deficiencies of water pollution source localization methods based on physical models in sensor networks,a centroid localization algorithm independent of diffusion models is proposed. In the algorithm,the area where the pollution source exists is determined firstly;the centroid of the area is calculated secondly. The posi-tion of the centroid is the source location. To solve the centroid localization problem,the particular solutions consid-ering the concentration contours are given. The experiment results show the advantages of the proposed localization methods comparing with the algorithms relying on diffusion models and the coarse localization algorithms. The effec-tiveness of the proposed algorithm is proved.