Research On Weak Beacon Light Signal Processing Of ATP System In Space Optical Communication

Space optical communication technology has the advantages of high data transmission rate and strong anti-interference energy,and is gradually applied to fields such as deep space communication and large-scale networking.This type of terminal uses the Acquisition,Tracking and Pointing(ATP)system to detect the angle of the incident beam and control the pointing of the optical antenna to establish a stable communication link.In order to reduce the power consumption and volume of optical communication terminals,the new generation ATP system is evolving to beam position detection technology without independent beacon light.This type of ATP system abandons the independent beacon laser,but separates part of the beam from the signal light as the beacon light,so that the beam energy received by the spot position detector is reduced.In addition,for flexible networking,the transmitting and receiving laser beams in the communication terminal are in the same waveband and share some optical components,which causes part of the local transmitting light to be reflected and enter the spot position detector together with the received light.In order to solve the problem of accurate position detection of weak beams,the paper chose a four-quadrant detector(QD)which has high sensitivity and can directly output continuous photocurrent signals.Research on the suppression of strong noise and the separation of strong interference signals from the perspective of line spectrum signal enhancement.The main work of the paper is as follows:(1)The QD detection model is studied,and its main noise distribution model is analyzed.QD detects the spot position according to the ratio of the amplitudes of the four output photocurrents,and its detection accuracy is affected by the radius of the spot,the spot position and the output signal-to-noise ratio(SNR).Therefore,the core of the QDbased spot position detection technology is how to improve the detection accuracy of the photocurrent signal amplitude value in a strong noise environment.The main noise that affects QD is background radiation noise,thermal noise and shot noise.These types of noise can be equivalent to broadband Gaussian noise.(2)The paper uses a single frequency signal to modulate the intensity of the beam,and converts the QD output signal into a line spectrum signal,thereby the accuracy of spot position detection can be improved with the method of line spectrum signal analysis.The energy of the line spectrum signal is concentrated on a certain frequency point,so the fast Fourier transform(FFT)can be used to obtain the amplitude value of the QD output photocurrent.When the output SNR is-7.83 d B,the absolute error of the spot position detection is maintained within 0.016 mm.However,this method does not suppress the noise,resulting in an excessively high noise floor.When the SNR drops to-17.86 d B,this method cannot provide sufficient SNR gain and cannot guarantee the detection accuracy of beam position.(3)The paper proposes windowed time reversal convolution(TRC)and cyclic crosscorrelation(CCR)line spectrum enhancement algorithms to suppress strong noise to obtain a larger SNR gain,and solves the problem of accurate position detection of weak spots under the condition of extremely low SNR.The TRC algorithm uses the autocorrelation characteristics of the line spectrum signal in the time domain to change the energy distribution state of the effective signal and noise,disperse the energy of the valid signal in the signal envelope,and concentrate the energy of the independent broadband noise signal in the signal at time 0.Then,the window function can be used to suppress noise and retain effective signals.So that,the signal line spectrum is enhanced,and the spectrum line amplitude of weak photocurrent signals accurately detected.However,this method sacrifices part of the effective signal energy while suppressing noise,and leads to a decrease in the linearity of the beam position detection result.Aiming at the problem of low linearity of the TRC algorithm detection results,the paper proposes the CCR algorithm to detect the amplitude of weak photocurrent signals.By constructing an infinite reference signal sequence,the algorithm can perform multiple cross-correlation operations with the detection signal repeatedly,and accumulate valid signals in the time domain,while strong uncorrelated noise cannot produce cumulative effects.Therefore,the CCR algorithm can suppress strong noise without loss of useful signal energy.In the experiment,both methods can accurately detect the spot position.When the QD output SNR is as low as-17.86 d B,the TRC algorithm can limit the absolute error within 0.0162 mm,and the CCR method can limit the absolute error within 0.018 mm.(4)Aiming at the interference of strong reflected light at the same wavelength,the paper uses polarization state isolation technology to suppress the reflected light,and the attenuation of the reflected light reaches about 63 d B.However,the polarization isolation technology cannot completely isolate the reflected light,and the reflected light power of the incident detector is 14 d B higher than that of the signal light power.The paper proposes to use the spectral line selection ability of the line spectrum enhancement algorithm to solve the problem of strong reflected light interference.the transmitting and receiving beams are modulated with narrowband signals of different frequencies,then the adaptive notch filter(ANF)is used to separate the line spectrum,and finally the TRC algorithm is used to enhance the valid signal spectrum,so as to detect the amplitude of the photocurrent.To solve the problem of the poor linearity of the ANF-TRC method,the paper derives the conditions for the CCR algorithm to suppress single-frequency interference signal,and constructs a reference signal according to the conditions,and successfully completes the separation of strong interference line spectrum signals.In the experiment,when the power of the two beams irradiated on the QD target surface differed by 100 times,both methods successfully completed the detection of the spot position,and the detection accuracy reached 0.0148 mm and 0.0165 mm,respectively.