基于支持度的多传感器一致可靠性融合

Consistent and Reliable Fusion of Multi-Sensor Based on Support Degree

对于任何先验知识未知的多传感器检测同一目标的情形,提出利用指数衰减函数度量传感器观测值的相互支持程度。根据支持度矩阵,获得各传感器在不同时刻的一致性度量。引入与一致性度量序列均值正相关,与方差负相关的序列作为传感器的权系数。用三个噪声分布未知的热电偶对恒温箱温度的观测值检验算法,仿真结果表明:由于既考虑了观测值的一致性,又考虑了传感器的可靠性,算法是有效的。

In the case of any prior knowledge is unknown, the exponential attenuation function was used to measure the mutual support degree of observation value of multi-sensor. Based on the support degree matrix, each sensor＇s consistency value of different time can be obtained. The time sequence, which is plus correlation with mean value of consistency value series and minus correlation with variance of consistency value series,is used as weight coefficient for sensor. To verify the effectiveness of this algorithm, three thermocouples, whose observation noise distribution is unknown, were used to detect the temperature of a constant temperature box. Taking both consistency and reliability of observation value into account, the simulation result shows this algorithm is effective.