Research On Wavelet De-noising Of Vibration Signal Of Vehicle-borne Precise Optical Measuring Instrument

The measuring basis of vehicle-borne precise optical measuring instrument will be affected by the vibration of the vehicle body,thus the measuring accuracy of the instrument is corrupted.In order to solve the problem,the vibration data generated by the engine and generator can be collected to compensate measurement result of the precise optical measuring instrument,and then high precision measurement results are achieved.In the process of collecting the vibration signals of the vehicle platform,it is often disturbed by some sampling noise.Therefore the vibration signals fail to reveal the practical vibration situation of the vehicle platform.It is necessary to eliminate the noise from the polluted vibration signal,and then use it to compensate the measurement data of the precise optical measuring instrument.The de-nosing process is of great significance to ensure the measurement accuracy of precise optical measuring instrument.The wavelet threshold de-noising algorithm is widely used among the commonly used signal de-noising algorithms.Four main factors of the algorithm are taken into account: wavelet basis function,decomposition level,and threshold function and threshold selection criterion.The threshold function and the threshold selection criterion are mainly discussed in this paper.In recent years,the threshold neural network capable of adaptive selecting threshold has been widely used in the field of wavelet threshold de-noising filed.Based on the wavelet threshold neural network algorithm,a novel threshold function is constructed as the excitation function of wavelet threshold neural network.The threshold value is chosen through supervised learning and non-supervised method separately,and the optimal de-nosing results are achieve.In order to verify the effectiveness of the algorithm,a de-noising flow is proposed based on the power spectral and energy distribution of the signal model.Using the sine signal model and sweep signal model,the de-noising effect is compared with the relevant de-noising algorithms.In order to verify the de-noising effect of the wavelet threshold neural network algorithm,a practical vehicle platform is built.The setup is composed of acceleration sensor,data collector,and computer software acquisition system and vehicle platform.The frequency distribution of the vehicle platform precise optical measuring instrument is obtained using power spectrum analysis method.Then the improved algorithm is used to de-noise the vibration signal of the precise optical measuring instrument.According to the criteria of SNR and RMSE,the proposed de-noising algorithm is compared with other signal de-noising algorithms.The experimental results show that the proposed algorithm in this paper is superior to other de-noising algorithms in eliminating the trend term noise and the random noise of the vibration signal.