Nanoscale Grating Displacement Measuring Key Technology

In contemporary industrial field such as microelectronics and ultra-precision manufacture, the displacement measurement with high precision is in great demand. Not only the measurement resolution is required to reach nanometer level in some application, but also the measurement distance is needed to get longer and longer. In this dissertation, the technologies are studied which correlate to grating displacement measurement system with long-range displacement and nanometer resolution.Fringe subdivision is the main and also the only approach for grating displacement measurement system which is required to reach nanometer level resolution. Software subdivision methods are always applied to it. The traditional subdivision methods are based on two ideal orthogonal sine and cosine signal. In many applications, high quality fringes signals can not be easily obtained, so the traditional methods are restricted.Three new fringe subdivision methods are first proposed opposing to the restrictions of the traditional subdivision methods. They are fringe subdivision method based on self correlation coefficient, fringe subdivision method based on gene algorithm and fringe subdivision method based on Fourier-transform. The principles and how-to-reality of the three methods are introduced respectively. Simulation and experiments have been done to test the validity of the three fringe subdivision methods.Fringe subdivision method based on self correlation coefficient has high processing speed.However it is inclined to be affected by harmonic wave, consequently it is more adaptive to deal with the signal with high quality. Fringe subdivision method based on genetic algorithm has slow processing speed, but it isn't easy to be affected by harmonic wave. So it is suited to be adapted when signal changes slowly. The processing speed of the fringe subdivision method based on Fourier transform is between that of the first method and the second method. But this method can be widely applied because it has no rigorous restriction to the quality of fringe signal. And it is the best choice among the three methods especially when the quality of the fringe signal is bad.