中文摘要：

火灾早期阴燃状态下，释放热量少，却产生大量的烟气，而风场中烟气的传播除了受自身的各向同性扩散影响外，更主要是受外界风的影响。基于斜风场中的气体浓度衰减模型，利用空间分布传感器对火灾源点进行定位，并提出了，一种加权距离差法来修正定位误差。该方法首先确定每个传感器对应的估计范围（集合）．然后通过寻求所有集合的交集来估计出火源点位置。结果表明，该方法不仅能提高定位精度，而且克服了迭代算法难收敛问题，具有实现简单，运行稳定的优点。

英文摘要：

This paper is aimed to propose an early fire source locating method based on the weighted differential distance approach in the skew wind field. In fire control, it proves very important to study the characteristic features of smolder when less heat was dissipated with more smoke released. The dispersal of the smoke is affected not only by isotropic diffusion in the wind situation, but also mainly by the wind. Based on the study of the attenuation model of smoke at the presence of the skew wind, this paper attempted to locate the fire source by spatially distributed sensors, with each sensor measuring the smoke concentration. The objective function of locating device consists of the measured value and the caleulated concentration by the attenuation model. When the objective function was set up, we have introduced an optimal algorithm into the method to minimize locating gap so as to be able to have a better approach to the fire source location. We have also proposed a weighted differential distance approach as a weighted factor of the distance information. In order to modify the locating error, first of all, we have to optimize the use of each sensor so as to locate the fire source by modifying the ohjection function, then locate the fire source position by intersecting all the sets. Simulation results show that this method can not only provide more accurate localization results, but also overcome the problem of poor convergence of iterative algorithm while implementing the estimation. Thus, it can be said that the proposed method is a practical locating means, efficient in providing some useful information for fire-extinguishing.