摘要
基于气体污染源浓度衰减模型,分别采用极大似然预估算法(M LE)、非线性最小二乘算法(NLS)对气体污染源定位进行了研究。仿真实验对比了两种算法在不同的传感器节点以及背景噪声情况下对预估定位误差的影响。结果表明:当环境背景噪声较小时,NLS可以得到比M LE算法更精确的预估结果。当环境背景噪声较大时,M LE算法比NLS算法有着更强的鲁棒性。
Based on the attenuation model of the plume, the location of plume source using maximum likelihood algorithm and the nonlinear least squares algorithm were studied. The effect of the estimation error, with different sensor number and different back ground noise, is researched by simulation. The result shows that better accuracy can be got by using nonlinear squares algorithm when the background noise is less. On the contrary, the maximum likelihood algorithm is robust to the much noise compared with the nonlinear squares algorithm.
出处
《华东理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2006年第7期780-783,共4页
Journal of East China University of Science and Technology
