无线传感器网络中基于贝叶斯技术的气体源定位研究

将无线传感器网络(WSN)引入气体源预估定位。基于贝叶斯原理，提出一种适用于WSN的气体污染源定位算法——改进粒子滤波(IPF)算法。该算法采用权重质心法确定预估定位初始点、退避定时排序法确定参与定位节点信息发送顺序及采用残差重抽样算法减少抽样方差以提高滤波性能。对比研究了改进粒子滤波( IPF)、扩展卡尔曼滤波(EKF)及改进的非线性最小二乘(I-NLS)算法定位性能，仿真表明构建的WSN对气体源定位有效，应用IPF算法定位性能优于EKF及I-NLS算法。

Plume Source Localization Based on Bayes Using Wireless Sensor Network

A wireless sensor network (WSN) was introduced into predicted evaluation of plume source localization. Based on the Bayes principle, an improved particle filter (IPF) algorithm was proposed, which is adapted for the gas pollution source localization using WSN. It was used to improve the filter performance that the methods of the weighted centraid, the backoff timer sorting and the residual re?sampling were adopted to determine the initial point of predicted location, the information sorting of located node and reduce the sampling variance respectively. The comparison of the simulated location performances among the IPF, the extended Kalman filter (EKF) and the improved non-linear least square (I-NLS) shows that WSN is effective on the plume source localization; IPF is better than EKF and I-NLS.