| Free Space Optical Communication(FSO)needs to establish a stable communication link to ensure communication quality,and the environment in the space is complex and changeable,so it highly requires the alignment of the terminals at both ends of the communication and the real-time tracking performance during the communication process.Aiming at this requirement,this parer will study on two aspects:laser spot position detection and tracking control technology,which ensure the stability of the FSO's communication link.The main research context is as follows:1.First,based on the analysis and comparison of the performance of the three commonly used photoelectric detectors in optical communications and combined with the experimental requirements,it is decided to use the Quadrant Detector(QD)in this paper.Then it introduced working principle of QD,analyzed the error source of causing the spot position detection error,and a band-pass filter is designed to remove the noise.Besides,the spot position detection system based on QD is designed.In order to improve the accuracy of spot position detection,the BP neural network position algorithm is proposed as the spot position detection algorithm in this paper.2.The back-end circuit of QD was implemented in hardware,and the software of FPGA data processing module was written in verilog language,and an experimental platform for spot position detection based on QD was built to collect experimental data.The collected experimental data will be trained by the designed neural network.Then,the most suitable configuration is selected by comparing the effects of different training functions and different transfer functions.Then the data of other intervals along the x-axis are collected again to verify the trained neural network structure.The overall error accuracy of the experiment in the x-axis direction of the QD can reach the level of10-4mm,the maximum error is only4×10-4 mm,which is very small and verifies the effectiveness and superiority of the neural network spot position detection algorithm.3.The environment in the laser communication process is complex and changeable.In order to improve the alignment accuracy of the laser communication system and deal with disturbances such as atmospheric turbulence and mechanical dither during the tracking process,the servo mechanism needs to be controlled by a controller to suppress disturbances after the deviation of the spot center is detected by QD.The factors affecting laser alignment and tracking are analyzed,and the tracking system plan is designed,and the adaptive control technology is proposed as the control technology of the system.Then it focuses on the theoretical basis of control methods such as digital PID control algorithm,neural network PID control algorithm,theoretical knowledge of neural network synovial control algorithm,and designs these control algorithms.4.Use MATLAB software to verify the control effect and anti-interference performance of several control algorithms designed,and simulate under three different disturbances.The simulation results denote that,compared with the traditional PID control algorithm,the neural network controller can adjust the weights adaptively in the face of complex and changeable disturbances,and the control accuracy of the neural network control algorithm can reach the level of 10-4mrad,and the dynamic response time can reach the level of 10-1s.In the face of disturbances,compared with the traditional PID control algorithm,the neural network controller has superior performance.Under the lager changes of sinusoidal perturbation,it can reach the control accuracy of1.49×10-4mrad,and the the dynamic response time is 0.2s.The tracking control performance meets the requirement of communication. |
PDF Full Text Request