Research Of An Optimal Filtering Method Based On LabVIEW

This paper presents a new filtering method to achieve a better filtering effecteasily, in the background of the rapid development of filtering technology, thesuperiority of wavelet filtering and the excellent features of virtual instrumentaltechnology.After studying on principles of three wavelet filtering methods and comparingtheir advantages and disadvantages, a novel filtering method is presented. It includesthree aspects as follows:Firstly, stationary wavelet transform is used in denoising. The decomposition ofnoised signal, base on translation invariant and redundancy of stationary wavelettransform,overcome the disadvantages of orthogonal wavelet transform and improvethe performance.And then, a novel wavelet denoising method by spatial correlation optimalthresholding is deduced. Different dependences among wavelet coefficients arestudied to denoise in the method. After significance measure based on spatialcorrelation is constructed, the procedure of wavelet shrinkage denoising bysignificance measure function is introduced. On this basis, through minimizing theGeneralized Cross Validation function,a optimal threshold which is obtained in theMean Square Error(MSE) sense,applies the optimal thresholding functiong based onspatial correlation, so that the algorithm of denoising by spatial correlationthresholding can remove the noise and preserve signal's original features effectively.Finally, a novel phase-locked loop (QPLL) is introduced. It has not only goodadvantages of large lock range, short locking time and high precision after locking,but also better characteristics of immunity with respect to noise and distortions. Itrealizes providing the input signal and optimum filtering by the introduction of QPLL,so that the novel denoising method in this paper is more innovative, and achieves abetter filtering effect.The experimental platform is constructed to prove this new filtering method,which includes two parts: FPGA signal acquisition and LabVIEW software filter. Thedata is received by FPGA from sensors and A/D transfer. The system tasks FPGA as CPU which controls A/D conversion, stores the data of the measured waveform andsends the storage data to the serial port when it receives the order from LabVIEW.Under the LabVIEW environment, the data needed is detected from serial port, andthen filtering processing is carried out by a novel wavelet method in backgroundgraphical program of the virtual instrument, while experimental results are shown inthe front panel.The results show that the novel filtering method not only can achieve a goodfiltering effect, but also be easy to use. Therefore it is proved to be an effectivefiltering method.