Physical Electronics New Type Of High Precision Grating Measurement System

The precision measuring instrument is a prerequisite for precision manufacturing,and grating measuring system is the research focus in precision measure field.It uses non-contact optical measurement,which is based on the machine,optics,micro-electronics and other new technologies,can convert the displacement of the measured object into the corresponding electrical signal.Grating measuring system is often used to measure the displacement,angle and other geometric measurements,which has advantages of the high precision,wide applicability,non-contact,no wear and so on.Due to the mature technology of miniaturization,laser emitter,focusing optical paths,photodetector and hardware are intergrated in the DVD pickup head.Therefore,the DVD pickup head is suitable to measuring the optical measurement.And the displacement measuring system based on the DVD pickup head has the advantages of high precision,non-contact,low cost etc.Which is ideal for measuring displacements.First,this paper studies the internal structure,working principle and focusing property of the DVD pickup head.In DVD pickup head,the displacement of focal plane with its focus error signal voltage is linear in a certain range.This paper uses the focusing characteristic to combine the DVD pickup head with the grating,and designs the grating measuring system with high integration level.The grating measuring system is build,assembled and debugged on the high precision vibration isolation platform.And the measure experiment was performed by the built measuring system.Secondly,this paper studies and compares the traditional methods of dynamic measurement error modeling and prediction methods.Support vector machine(SVM)is often used to predict the measuring system's dynamic measurement error.The key parameters of SVM were always selected by optimization algorithm.But the traditional optimization algorithm is fall into the local optimal easily.Which can not ensure the model's performance.In this paper,an improved particle swarm optimization(NAPSO)is proposed,which is based on simulated annealing mechanism and natural selection method.Natural selection and Simulated annealing are added in particle swarm optimization(PSO)to raise the ability to avoid local optimum.These two operations are used to both ensure the convergence rate of the algorithm and guarantee that the ability of the algorithm to jump out of the local optimal trap can be enhanced.And the operations can compensates for each other,the simulated annealing operation can increase sample diversity,and the natural selection operation can accelerates the convergence rate.To verify the performance of NAPSO,NAPSO and other four improved PSO algorithms are used to optimized five standard test functions.The comparison results demonstrates the ability of the NAPSO.Then,the NAPSO is used to select the key parameters of SVM and build the prediction model.Three prediction models are used to predict the built system's dynamic measurement errors,they are the PSO-SVM,the NAPSO-SVM,and the glowworm swarm optimization SVM(GSO-SVM).The root mean squared error and mean absoluter percentage error are employed to evaluate the prediction models' performances.The experiment results show that the NAPSO-SVM has a better prediction precision and a less prediction errors among the three prediction models.The NAPSO-SVM prediction model provides the theoretical basis for the error correction method of real high-precision grating measuring system.At last,the prediction results are used to correct the measuring system's measurement results,the system's measurement precision is improved through this correction method.