Study On Adaptive Subdivision Method And Error Correction Algorithm For Grating Signal

Grating sensor not only has the advantage of high measuring range and high accuracy,but also is easy to digitalization in industrial control system,so it has been widely applied in the field of precision automatic measurement.But whether it is absolute grating or incremental grating,the principle is the sensor will output a period signal,if the object measured move a unit grating pitch.With the analysis of the output signal,the moving displacement of the object can be obtained.So,the accuracy of the grating line determines the accuracy of the grating sensor,the density of the grid line determines the resolution of the grating sensor.And because the restrictions of manufacturing industry,the density of the grid line is limited.Therefore,it is one of the main methods to improve the resolution of the grating sensor by auxiliary subdivision for grating's original signal.After a thorough study for the grating sensor' characteristics of mechanical and motion,a prediction subdivision model is proposed in this paper.According to the difference of the motion state of the grating,the model can realize the adaptive subdivision for the grating signal.The method uses high-speed embedded devices to sample the signal output from the grating,obtains the passing time of the historical unit grating pitch and compose the observation sample sequence.The adaptive predictive algorithm is used to construct the predictive mathematical model and correct the prediction error in real time.then predict the running time of the next grating pitch.Before the raster moves to the next raster line,the embedded system synchronously output subdivided pulses which represents the subdivision displacement,completing subdivision of the raster signal.Different from the existing subdivision methods,the subdivision effect of the prediction subdivision model has nothing to do with the sinusoidal characteristics and amplitude characteristics of the grating output signal.In this paper,a round grating ROD880 produced by HEIDENHAIN is used as a subdivision object,and the following contents are studied.(1)The traditional displacement measurement subdivision method and the time-based subdivision method were introduced.The advantages and disadvantages of using two models for grating signal subdivision are analyzed.Finally,the predictive subdivision method of grating signal is proposed.(2)The collected sample series is analyzed to divide the motion status of the grating sensor.The second exponential smoothing method was used to build the predict model when the grating motion was steady.When the grating is in the "critical" state,the gray prediction algorithm is used to build the prediction model.To improve the prediction accuracy,a Markov error correction model is constructed to correct the prediction error.(3)Using high-speed ARM + FPGA architecture to achieve the embedded hardware platform design of the subdivision system.The experimental research on the subdivision ability of the grating under different motion conditions was carried out.At the same time,the factors affecting the subdivision effect were analyzed.The dynamic experiment of the grating under different motion conditions is carried out.,The analysis of the experimental results shows that using the adaptive subdivision method of the grating signal and error correction algorithm proposed in this paper,when the subdivision multiple is set to 400,the subdivision error within the unit grating pitch is ±0.65?.