一种基于多方法的多传感器数据融合算法研究

提出了一种基于多方法的多传感器数据融合算法,分批估计融合求得单传感器最优估计值,然后依据权值最优分配原则进行分组自适应加权融合处理.通过对多热电偶测温的实测数据分析表明,与算术平均值法、单传感器分批估计和自适应加权相结合的算法以及单传感器分批估计和多传感器分批估计相结合的算法相比,绝对误差分别降低3.52℃,1.28℃和1.227℃,相对误差分别降低0.294％,0.107％和0.102％.

Study of Multi-sensor Data Fusion Algorithm Based on Diverse Methods

In this paper,a kind of multi-sensor data fusion method based on various methods is put forward.The optimal estimation of single sensor can be completed by batch estimation fusion.Then grouping adaptive weighted fusion is adopted in processing the optimal estimations of these sensors,based on the optimal distribution principle of weighted value.According to the analysis of experimental data,the absolute error is respectively reduced by 3.52℃,1.28℃ and 1.227℃ and the relative error is decreased by 0.294％,0.107％ and 0.102％,compared with the methods of arithmetic average,the algorithms which consist of the batch estimation fusion of single sensor and adaptive weighted fusion of sensors and combine the batch estimation fusion of single sensor and that of sensors.