摘要
传感器信号的非线性程度直接影响测量系统的精度,因而,必须对其进行线性化。针对传感器特性和转换电路的非线性可以采用硬件的方法进行线性化;而对其他因素的影响,可以采用智能系统,用最小二乘法进行曲线拟合、神经网络法数据融合等软件编程来实现;或者根据具体问题将二者结合。实验证明:这些都是实现线性化的好措施。
The sensor signal non-linear degree directly affects the entire measurement system precision,therefore must be made linearization. Aimed at sensor characteristic and switching circuit linearization, the hardware measure may be taken to complete the linearization; the intelligent test system may be used for other factor influence. The curve fitting carried on with the least squares method, software programming and so on nerve network method data fusion are realized ; or basing problem combines them ; the experiments prove all are good measures to realize the lining.
出处
《传感器与微系统》
CSCD
北大核心
2008年第12期36-38,共3页
Transducer and Microsystem Technologies
基金
陕西省自然科学基金资助项目(2004E202)
教育部春晖计划资助项目(Z2005-1-61004)
关键词
传感器
线性化
曲线拟合
神经网络
sensor
linearization
curve fitting
neural network