摘要
针对采用基于电容传感原理的加速度传感器来实现角度测量时出现的非线性问题,提出数据预处理与广义回归神经网络(Generalized Regression Neural Network,GRNN)相结合的非线性标定技术.结合实验,详细分析该技术的标定过程,并对该技术与单一使用GRNN进行标定的标定结果进行对比.实验结果表明,该技术标定速度快、精度高、鲁棒性强、实时性好,能有效拓展电容式加速度传感器用于角度测量的线性范围,有较高的工程实用价值.
When capacitive accelerometer is used in angle measurement, the measurement results are inaccurate because of its non-linear characteristics. In order to solve this problem, a non-linear calibration technology, which integrates the data pre-process and Generalized Regression Neural Network （GRNN） , is put forward. The calibration process of this technology is analyzed in detail through experiments. The calibration results and the other one, which only uses GRNN to calibrate, are also compared in the experiments. The experiments show that this technology has the features of fast calibrating speed, high precision, strong robustness and good real-time performance, which greatly expands the linear range of the capacitive accelerometer for angle measurement and has great value in engineering measurement.
出处
《上海海事大学学报》
北大核心
2008年第3期37-40,共4页
Journal of Shanghai Maritime University
关键词
电容式加速度传感器
非线性标定
广义回归神经网络
角度测量
capacitive accelerometer
non-linear calibration
generalized regression neural network
angle measurement
