Comparative Study Of Absolute Gravimeter Laser Interference Signal Processing Algorithms

Objects on earth are affected by gravity,the Earth's gravitational sphere of influence are called Earth's gravitational field.Earth's gravitational field can reflect the internal structure and physical changes of the Earth's motion.Gravity field research and observation can be widely applied to various areas such as seismic research,resources exploration,construction,control and guide and so on.Absolute gravimeter can accurately measure the acceleration of gravity,thereby measuring the distribution of the Earth's surface gravity field.It has important applications in the field of earth sciences and measurement.First,falling body falling through a vacuum chamber to generate a laser interference fringes.Then convert optical signals into electrical signals and extraction"displacement-time"from the voltage signal.Finally,the results or g obtained by linear or non-linear fit.Using the appropriate algorithm can meet the accuracy requirements and improve measurement speed.This paper describes a mathematical model of absolute gravimeter laser interferometer signal and then three common signal processing algorithms are described,that is double sampling zero-crossing detection method,windowed second order difference method and complex heterodyne method.The main work is as follows:1.Using MATLAB software simulate absolute gravimeter laser interferometer signal,the frequency of the signal simulated increased from DC to 6MHz in 0.2 s,which is similar to the actual laser interferometer signal.2.Without any interference added in the signal,using double sampling zero-crossing detection method,windowed second order difference method and complex heterodyne method respectively to get g value,deviation,time-consuming and residuals.All the three methods can reach 0.01?Gal magnitude.Fit residuals of zero-crossing detection method and windowed second order difference method can reach 10^-10 m magnitude,the fit residuals of complex heterodyne method can reach10^-9 m magnitude.3.2%amplitude noise is added to the signal,then using three methods respectively to deal with the signal.As can be seen from the results,even without any filtering process,deviation of this methods are not more than 0.1?Gal magnitude.Fit residuals of double sampling zero-crossing detection method increased to 10^-9 m magnitude,fit residuals of windowed second order difference method can still reach10^-10 m magnitude.Fit residuals of complex heterodyne method can reach 0.02 m magnitude if no filtering process added.From the deviation,fit residuals,time-consuming and sampling rate aspect,three methods are compared,the conclusion is:nowadays hardware like data acquisition card and computer is no longer a limiting factor,when required deviation does not exceed 0.1 uGal magnitude and fit residuals less than 10^-8 m magnitude,all the three algorithms can meet the requirements.But in order to increase the processing speed and convenient for following correction processing,double sampling zero-crossing detection method is the most appropriate algorithm currently.