能抵抗粗差的雷电定位算法研究与仿真

雷电是严重危害人类生命财产安全的自然灾害。在雷电监测系统中，定位计算直接关系到探测结果的精度。然而在探测到的原始数据中有许多包含粗差，粗差的影响使得由基本定位方法得到的结果严重偏离真实值。为了满足应用的要求，必须设计能够抵抗粗差干扰的定位方法。论文首先介绍了目前在用的三站定位方法,并严格推导了Taylor级数法。为了使基本定位方法能够具备抵抗粗差的能力，基于数据挖掘技术设计了两种粗差处理算法：k-means聚类法和决策树分类法。仿真说明,采用后两种方法能够有效地抵抗粗差的干扰,提高定位精度。

Algorithms and Simulation for Robust Lightning Location

Lightning is a natural disaster for human being. In lightning detection and location system, location algorithm is very important. But many original data have some gross error, which can make the result far depart from the real location. So in order to satisfy the requirement of application, more location algorithm which can resist the gross error should be developed. This issue firstly introduces the basic method of lightning location, and then gives a strict reduction for Taylor method. In order to advance the precision of lightning location, two data mining methods are designed for lightning location: k-means clustering method and decision tree classification method. Simulations show that the new algorithms can control gross error more efficiently and hence enhance the location precision.