The Application Research Of Signal Processing For Laser Doppler Velocimeter

The Laser Doppler Velocimeter (LDV) is widely used in the science research and industry fields due to its many technological advantages, such as non-contact measurement, linearity, high resolution and rapid dynamic response, etc. But the Doppler signal possess the characteristics of heavy noise interference, broad frequency change and discontinuous signal, its analysis and processing affects directly whether the whole system is better or not. Thus, removing the noise effectively and improving the measurement accuracy become a significant question for the LDV system.A corresponding signal processing method is proposed in this paper according to the characteristics of Doppler signal. At first, in order to decrease the noise interference, a self-adapted filtering algorithm is introduced. On the basis of LMS algorithm, a novel LMS with varying step distance is presented. This vary step distance factor not only measures the detailed changing process of the noise variance and the error signal, but also needs relatively less computation, allows this improved algorithm to have better convergence property and filter effect. In addition, since the signal has a kind of dense spectrum, a re-sampled algorithm that can chose zooming multiple self-adapted according to frequency band is researched, is an improvement of traditional Zoom-FFT. For the frequency band of interest, after choosing the zooming multiple, is approached real frequency. This algorithm solves the problem of the traditional compound modulating that constrains the computing velocity, increases the high-frequency resolution and satisfies the demand of quickly zooming. Through the simulation, the frequency can be obtained accurately, and the high efficiency and robustness of proposed algorithm can be validated.