The Application Of Stochastic Resonance And Time-Domain Smoothing-Processing In Gravitational Experiments

In gravitational experiments, such as the measurement of the gravitational constant G, the experimental test of Newtonian inverse squared law, the weak Equivalent Principle experiment, and gravitational wave detection and so on. In general, the gravitational effects are too weak to submerge into the noises. The noises can be classified two kinds, one is the external interference source, and another is the internal noise. The external disturbances due to electromagnetic force, the change of ambient temperature of the laboratory, and seismic oscillation can be reduced by some ways, while the intrinsic noise which gives the basic limitations to the experimental measurement can never be eliminated. Generally, the noise does harm to the measurement of signal, but it can also play an active part under a certain condition. We can improve the signal-to-noise ratio of the system output by adjusting the parameters or increasing the strength of noise from a nonlinear device of stochastic resonance. I did some work on the noise for my master degree as follows: (1) Based on the extensive investigation to the noise of various types, the physical properties of the thermal noise, low frequency noise and shot noise are briefly summarized. (2) I introduce the mechanism of stochastic resonance in detail and apply this technique to detect the weak periodic signal immerged in heavy color noise. And this method is introduced preliminarily to some null experimental tests such as the limit on the photon rest mass with rotating torsion balance. (3) The effect of the time-domain smoothing-processing method for the equivalence principle test using free-falling, compared to the low-pass-filter method, is discussed. Theoretical calculation shows the time-domain smoothing-processing can suppress the high-frequency noises and does not cause pseudo acceleration. The physical reason is that the time-domain smooth-processing method is an unweighted filtering and unrelated to the polynomial fitting. The uncertain acceleration difference due to high-frequency mechanical vibrations for equivalence principle test for extended rotating bodies can be suppressed from 2 μGal to 0.05 μGal after the time-domain smoothing-processing.