小波去噪用于电磁感应式角加速度传感器测量

Electromagnetic Induction Angular Acceleration Sensor Measurement Based on Wavelet De-Noising

自行设计了一种电磁感应式角加速度传感器,阐述了传感器的基本机械结构和工作原理,然后采用该传感器对空载时的单相异步电机进行了实际的角加速度测试,验证了测量原理的正确性,并由实验波形观测到传感器的输出信号存在大量噪声,因此构建了一种基于小波变换的去噪方案。首先对传感器的实际输出信号采用sym8小波进行分解,然后利用半软阈值对含噪信号进行处理,最后对信号进行重构。结果表明该方法可以去除大部分高频噪声,能够恢复数据的真实性,进而提高数据的置信度和泛化性能,且去噪效果远好于传统的傅里叶变换。

A electromagnetic induction angular acceleration is presented, the basic mechanical structure and working principle of sensor are introduced, and its measurement principle is verified according to the actual test of the single-phase asynchronous when no load. While the sensor output signal is annihilated by lots of noise, a de-noising scheme based on wavelet transform is constructed. The actual signal of sensor is decom- posed with wavelet Sym8, then processed by half soft threshold, and reconstructed signal finally. Simulation results indicated that the method can get rid of most of the high frequency noise, recover the factuality and improve the fitting and generalization capability of the data, and de-noising effect is far better than the traditional Fourier transformation.