Research On Dynamic Error Correction Method Of Time Grating Displacement Sensor

As a new type of high precision displacement measurement sensor,time gating to the " Space time conversion " theory as design concept,reduced the precision mechanical reticle,improved the sensor's precision and reduced the cost.Found in the study of dynamic measurement accuracy is far better than the static measurement.Based on this,aiming at the dynamic measurement of time grating,proposed a feasible scheme of dynamic error correction,to further improve the accuracy of dynamic measurement of time grating.The main contents and conclusions of this paper shows as follows: 1)The theory of space time coordinate transformation and the design idea of the time grating angular displacement sensor is described,and the common methods of dynamic error compensation are summarized.2)The work principle and application of the three methods of error analysis(fourier transform,power spectrum density estimation,Allan variance)are compared and analyzed.The fourier transform(also called harmonic fitting method)is used to eliminate the system error of the time grating measurement,and the random error data of the time grating is obtained.According to Allan variance,the random error of time grating is divided into six categories: quantization noise,angle random walk,rate random walk,rate ramp,zero bias instability and sinusoidal noise.Then analyzed the mechanism of the six types error and effect on the measuring result.At last the least square method is used to fit the characteristic coefficients of these 6 kinds of random errors,so the main source of the random error of the time grating is determined.3)Comparative analyzed three most commonly used stationary time series models: auto Regressive Model(AR),moving average model(MA),auto regressive moving average model(ARMA).And the random error of time grating data time series modeling is established.Through model test,model identification,model order and parameter estimation,determined the time grating random error time series model-ARMA(2,3)model,prepared for filtering.4)The work principle of the conventional Kalman filter and the adaptive Kalman filter are described respectively,and the two algorithms are realized by MATLAB programming.According to the time series model established above,the state equation and measurement equation of the time gating of random error is obtained,then carried on the Kalman filter.Compared the effect of these two kinds of Kalman filter by the Allan variance method.Experimental and simulation results showed that adaptive Kalman filtering is better than the conventional Kalman filter.Adaptive Kalman filtering can effectively decrease random error of the time grating and improve the measurement accuracy of time grating.