A Novel Signal Processing Method For Azimuth-Gradienter Based On Mems

R Zhu,Zy Zhou
2001-01-01
Abstract:This paper discusses a new signal processing method for reducing the existing errors and getting stable output for the azimuth-gradienter based on micro electromechanical system (MEMS). The azimuth-gradienter is composed of magnetometers and accelerometers. The stochastic noises, drifts and instability come from a variety of reasons, such as magnetic field fluctuations, electrical fluctuations, and external disturbance, which are inevitable in magnetometers and accelerometers. Considering about these influences, an applied and effective signal processing method has been developed. Firstly, we design a self-adaptive non-linear filter aiming at the characteristics of the measured signals. The original measured angle and its forward shifts are as the inputs of the filter, and the output is as the estimation of the real heading angle. By properly evaluating and updating the parameters for the filter, its output will surely approach to the real value with less error compared with the original measure. Secondly, integrating the errors and the increments of errors in criterion function and minimizing it propose an advanced algorithm. The algorithm is used to perform the parameter teaming of the filter. The application results show that the above design provides a fast and effective way for eliminating noises and stabilizing the angle output.
What problem does this paper attempt to address?