Research On Real-time Processing Of The Signal Of Micro-displacement Sensor For Active Optical System

Active optical technology can correct the deformation caused by environmental factors such as gravity and temperature in astronomical telescope and the precision measurement of micro-displacement between the lens of the active optical system is the basis of the automatic adjustment.At present,inductive micro-displacement sensors used in active optics in China have been unable to meet growing demand of the larger large-scale telescope manufacturing in the future.Therefore,proposed for a new type of capacitive micro-displacement sensor,realtime signal processing methods are studied and discussed.Based on the analysis of the characteristics of the input signal,the digital filtering and transmitting modules of real-time digital signal received from AD converter are designed as a whole.A multi-channel digital filter bank based on a wavelet transform is designed to remove noise and gains a clean output signal.The simulation results of multi-channel wavelet filtering,mean filtering,median filtering and adaptive filtering are compared and analyzed,which have proved that the algorithm described in this thesis achieves good noise suppression capability.In this thesis,multi-channel wavelet is used as a tool to establish multi-channel filter banks and the multiphase matrix is used to reduce the computational redundancy of the filter.Through the studying of lifting algorithm,an improved lifting algorithm is used to make all elements in the filter matrix become integers,and the feasibility of hardware implementation is increased.The multi-channel wavelet filter can decompose the signal into four different frequency domains for processing.If hardware available,setting up a higher decomposition level can improve the smoothness of signal and reduce the nonlinear error of the sensor,which has strong flexibility.After completing the front-end algorithm development,multi-channel wavelet filter,mean filter,median filter and adaptive filter are implemented based on FPGA.Voltage conversion,digital filtering and data transmission of digital signals are completed on the Development Board of Xinlinx's XC6SLX16 which belongs to the Spartan-6 Series.The GUI interface is also designed based on MATLAB to receive and observe the data returned by the digital signal processing module.The results shows that the digital signal processing method described in this paper achieves good noise processing capability.