Mems Gyroscope Signal Dsp-based Noise Reduction Technology And Realization

MEMS gyroscopes can sense the angle rate of moving vehicle. Comparing with conventional electro-mechanical gyroscope and electro-optical gyroscope, MEMS gyro has the advantages of low cost, small size, low weight and high reliability .In the past decades MEMS gyroscope has become the major subject of widely research and development and will continue to boost the evolution of navigation technologies. At present, basing on the process level, MEMS gyroscope is restricted in the field of low level navigation.Drift noise is an important factor that affects the precision of gyroscope, how to decrease the random drift noise is one of the ways to improve the performance of the MEMS gyroscope, so the research on gyro drift filtering and modeling is much more important. In this paper, we analyze the signal of MEMS gyroscope in time domain and frequency domain, a Chebiv low-pass filter has been employed on the noise with frequency beyond passed-band, but low-pass filter has a big disadvantage ,the window domino offect is serious, then we design a forecast filter based on FIR ,it can keep linearity phasic and the simulation result is obvious, an improved FIR mean hybrid (FMH) filter is designed to filter the gauss noise in the pass-band, the result based on DSP indicated that this method can effectively improve the accuracy of MEMS gyroscope and keep filtered signal track in time, the mathematical morphology filter is designed to filter the ramp noise and pulse noise in gyroscope, it is very easy for its hardware implementation, kalman filter is the tradition method in navigation system, after the residual signal is modeled by a 1-order AR ,a classical kalman filtering algorithm is used to filter the signal, the accuracy of drift is improved greatly, the processed result form the practical gyroscope measurement shows the method of the de-drift is effective.We design a DSP-based signal de-noise system, TMS320F2812, the new DSP chip produced by TI company, is used in this system. Every algorithm can run well and this system makes a good effect in practical appliance because of its simple structure, small bulk, easy to integration and real time.