摘要
在充分研究和比较多种降低多传感器交叉敏感方法的基础上,介绍一种用于消除或降低压阻式压力传感器交叉敏感的张量积再生小波核支持向量回归的方法。这种支持向量回归核函数的构建是基于一种算子理论构造的再生小波核Hilbert空间的方法。实验结果表明,该方法可以明显地降低多传感器的交叉敏感。经其融合后,传感器的温度灵敏度系数αs和电流影响系数αI比融合前分别降低了近2个数量级,零位温度系数α0下降了约3/4。
On the basis of sufficient study and comparison of a variety of methods for reducing cross-sensitivity of multi-sensors, this paper introduces a SVR method of reproducing tensor product wavelet kernel, which can be used in eliminating or reducing the cross-sensitivity of piezoresistive pressure sensors. The building of SVR kernel function is based on a method in which the reproducing wavelet kernel Hilbert space is constructed by an op- erator theory. Experimental results indicate that the method can reduce effectively the cross-sensitivity of multi-sensors. The temperature coefficient （FSO TC） and the electrical current influence coefficient of the pressure sensor sensitivity are reduced nearly two orders respectively compared with those before data fusion, and the temperature coefficient of zero position is three quarters lower.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2009年第12期2557-2561,共5页
Chinese Journal of Scientific Instrument
关键词
多传感器
交叉敏感
再生核希尔伯特空间
支持向量回归
再生小波核
multi-sensors
cross-sensitivity
reproducing kernel Hilbert space
support vector regression
reproducing wavelet kernel
