基于参数估计的多传感器数据融合算法研究

The Study on Data Fusion Algorithms of Multi-Sensor Based on Parameter-Estimation

为提高多传感器数据融合精度,结合传感器网络的特点及应用实例,研究分批估计、自适应加权和方差估计算法在数据融合中的有效性、准确度和实时性,提出按测量方差的自适应加权数据融合算法,利用600个传感器所提供的实例数据,对几种算法进行仿真,并比较了几种算法的有效性及其融合精度的差异,其结果表明采用自适应加权算法可以有效提高融合精度,对考虑了环境噪声的多传感器数据采集系统较为适合.

In order to advance accuracy of multi-sensor data fusion, the validity and degree of accuracy and actual time of batched-estimation algorithm, weighing-estimation algorithm and variance-estimation algorithm were studied on multi-sensor data fusion, integrating the characteristic of the sensors network and its applications. Weighing-estimation algorithm by inspected-variance was proposed and the validity and their difference of the precision in data fusion were illustrated on PC simulation, basing the data from 600 sensors. The PC simulation shows that weighing-estimation algorithm by inspected-variance improves the precision efficiently, and it is more reasonable for multi-sensors data acquisition systems when consider environment noises.