摘要在现实生活和生产中,我们在实际的工业生产和学术研究中,不可避免的需 要用到各种各样的传感器。有一些传感器因为其物理构造等等原因是线性的,而 绝大多数的传感器的特性曲线都是非线性的。而电容式位移传感器的非线性的特 性成为了阻挡我们有效利用其性能的拦路虎。这个时候就需要我们对传感器进行 非线性的校正。83565
通常对传感器校正,大多数使用到硬件校正法和软件校正法。硬件校正法因其 难以实现,本身用到的元器件就有存在误差。而且衍生出的误差较大,并不推荐 使用。而随着软件的迅速的发展,软件校正法逐渐展露头角。一般会采用端基法 和最小二乘法进行校正。但考虑到端基法的误差较大,而最小二乘法的拟合效果 也不是很理想(误差的平方和不能控制在一个理想范围内)。最后我们引入了 BP 神经网络进行传感器的非线性校正。因为这么几种软件校正的方法均有他们 的局限性。所以相应的在后续,我们也会通过对 SVM 支持向量机,包括 LS-SVM 的调研,来对电容式位移传感器的非线性校正进行一个全方面的补充。
毕业论文关键词:电容式位移传感器,非线性校正,最小二乘法,BP 神经网络。
AbstractIn life and production, we need to use a variety of sensors in the actual industrial production and academic research。Some sensors are linear because of their physical structure and so on , the vast majority of the characteristic curve of the sensor are nonlinear。The characteristics of nonlinear capacitive displacement sensor has become the effective use of blocking our way。 At this time, we need to make a nonlinear correction of the sensor。
Usually ,the sensor calibration mostly takes hardware correction and software correction 。Hardware correction is difficult to achieve, there is an error in their own use of the components。 And the error is hard to change, so it is not recommended to use。With the development of software rapidly, software correction has gradually emerged。The general will be calibrated by terminal group method and least square method。 But taking into account the terminal group analysis method of the error is large, and least squares fitting effect is not is ideal (the sum of squares of errors can not be controlled in an ideal range)。However,we introduce the BP neural network to carry out the nonlinear correction of the sensor。 Because so several software calibration methods have their limitations。 So the corresponding in the follow-up, we will also through the SVM support vector machines, including the LS-SVM research, to the capacitive displacement sensor for a full complement of nonlinear correction。
Keywords: capacitive displacement sensor, nonlinear calibration, least squares, BP neural network。
目录
第一章 绪论 1
1。1 传感器校正的背景 1
1。2 传感器的一般特性 1
1。3 传感器的静态特性的主要指标 1
1。4 电容式位移传感器 3
第二章 校正方法 6
2。1 硬件补偿法 6
2。1。1 折线逼近法 6
2。2 软件校正法 6
2。2。1 计算法 7
2。2。2 查表法 7
2。2。3 插值法 7
2。2。4 端基法 7
2。2。5 最小二乘法