Research On Weak Signal Detection Method And Its Application

Weak signal measurement have a very wide range of application in the laser ranging,mechanical fault diagnosis,communication and other fields.It is the hotspot and difficulty of modern signal processing to detect the weak signal submerged in the background of noise.Based on the characteristics of weak signal and noise,this paper studied the Cross-correlation algorithm,the spectral analysis algorithm and the chaos algorithm based on Duffing oscillator.Firstly,This paper introduced the Cross-correlation algorithm and spectral analysis for the application of weak signal detection in Laser Ranging.Where,Cross-correlation algorithm extracts the weak signal from noise using the irrelevantity of weak signal and noise,then determines the delay time.Through the frequency-transformation,the spectral analysis algorithm Calculate the phase angles of the main spectral line for the weak signal and reference signal,and performing Subtraction operation,and then calculated the delay time.At the same time,this paper described the chaotic algorithm based on Duffing oscillator,according to the chaos characteristic that immunity to noise and the Sensitivity to periodic signals,established a weak signal detection model.Secondly,based on Matlab,this paper designed the phase detection model of Cross-correlation and spectral analysis.The paper obtained the selection criteria of the sampling frequency and data length for the two algorithms in accordance with different conditions.The frequency offset has a great influence to the two algorithms,when the frequency have a slight offset,the error more than 3%.Under noisy condition,although the SNR can reach-2dB for the Cross-correlation algorithm,but its performance is not stable;While the spectral analysis is stable,when the SNR is greater than 18dB,the maximum error is less than 3%.So paper using the spectral analysis do further experiment.Based on QuartusII it designed the phase detection model of spectral analysis,selected the sampling frequency which is 128kHz,and the signal frequency which is 1kHz,and the data length is 128 to simulate which used Modelsim,for the different phase angle conditions,the error is less than 0.08%which is 0.072°,Which laid a foundation for FPGA implementation.Thirdly,this paper built a hardware platform,finished the experiment of phase shift laser rangefinder based on FPGA,selected the system clock which is 25MHz,and the signal frequency which is 195312.5Hz,and the data length which is 128 to conduct experiment.Under the different phase angle conditions,the error is less than 0.09%,and the accuracy is 0.0234m.That is designed a high-speed,high-precision laser ranging method,which can be used in the design of laser rangefinder in engineering application.Finally,for the phenomenon of roller eccentricity,this paper introduced the model of roller eccentricity,and the eccentricity signal is weak periodic signal,so it selected the chaos algorithm for the detection,and took an roller eccentricity signal as an example,it verified the feasibility of the algorithm.According to the three problems of Duffing oscillator chaos model,it put forward the idea of combining variable scale and reverse transformation,improved the Chaotic Model of Duffing Oscillator,and made a preliminary analysis,which can effectively solve the problem.