摘要
针对已有降水传感器对强降水测量误差大、测量范围小等不足,研制了一种基于压力敏感元件的雨量智能传感器。设计的传感器硬件由压力敏感元件核心的降雨物理量转换单元、以差分放大和线性处理模拟电路为主的信号处理单元、以高速低功耗嵌入式处理器ARM9为核心的数据计算单元等部分组成;传感器软件上采用广义回归神经网络的函数逼近方法,通过小样本训练建立精确测量模型。测量降雨作用在压力传感器上的压力大小及其对时间的变化,利用广义回归神经网络模型可计算得到精确雨量。试验表明,该传感器不仅解决了强降水等原因引起的雨量测量不准的问题,具有测量量程宽、精度高等优点。
In order to overcome the disadvantages of existing rain gauges including the obvious inaccuracy while rainfall intensity is strong, and their measurement range is small, a novel automatic rainfall gauge based on the sensitive pressure sensing component has been developed. The hardware of the developed gauge includes a rainfall signal transformer based on a sensitive pressure sensor, a electrical signal processing unit built up with differential amplifier and linear transform circuit,and a computation and procession centre based on the higher speed and lower power consumption embedded processor ARM. The generalized regression neural network（Gt/NN） is applied in the software of the developed gauge to realize the approximation of function, and a accuracy measurement model has been built by training the GRNN with small data sample. By measuring the pressure and the duration response generated by the rainfall on the sensitive pressure sensor, and calculating with the trained GRNN model, the accuracy rainfall could be obtained. The test results show that the developed rainfall gauge not only could be applied to measure the heavy rain, but also has the merits of wider measurement range and more precision.
出处
《传感技术学报》
CAS
CSCD
北大核心
2012年第6期761-765,共5页
Chinese Journal of Sensors and Actuators
基金
山东省自然基金项目(ZR2010DL003)
关键词
传感器技术
压力敏感元件
雨量计
广义回归神经网络
sensor technology
pressure sensing component
rainfall gauge
generalized regression neural network
