The Error Analysis And Filtering Methods Research Of HRG

The Hemispherical Resonator Gyros (HRG) is a new type of gyro invented in 1960s. Compared with traditional mechanical and optic gyros, HRG has such advantages as no high speed circumvolving or moving parts in structure, no warm-up time and short start-up time, wide signal band in frequency, low excursion noises, great endurance in over loading, nuclear radiation and short time power off, small bulk, light weight, low power cost and long life, which is suitable for space applications. This paper mainly discusses the errors analysis of a national type of HRG, and research on its signal filtering based on the signal characteristic. A dynamic angle measuring experiment is carried on after theory analysis. A simulation of its application feasibility in moon vehicle is also discussed.On the aspect of errors analysis of HRG, a brief introduction of the close-loop detection theory of HRG is present first. Later the expressions of close-loop detection errors are deduced in theory. The impact of stimulating signals, turning speed, outside appended oscillation, the location & area of electrodes, and the current loss in capacitor is considered, and the primary analysis of error compensation is given. The Allan Variance (AV) analysis method is used to evaluate the HRG signal. The result indicates that the main random error of HRG is Angle Random Walk (ARW). In the end of this chapter, the Power Spectrum Density (PSD) analysis of HRG is given, which shows that the main noise of HRG is low frequency noise.On the aspect of signal filtering of HRG, the Forward Linear Prediction (FLP) filtering method is introduced first. FLP is a kind of real time filtering method which is capable of reducing the ARW of gyros efficiently. The wavelet transform filter and infinite impulse response (IIR) digital filter is applied, and the output signals of the three filters are compared. It is concluded that FLP is better than other two filters in reducing the ARW, constraining the high frequency noise and reserving the low frequency useful signals. Another advantage of FLP filter is that it can build error model at each startup of HRG, which has advantage in the case that the error model changes at the startups. It is shown that the FLP filtered signal can track the change of the output signal of HRG. And the precision of the former is better than the later.On the aspect of its application feasibility in moon vehicle, a kind of orientation and positioning system based on HRG is discussed. Simulation results of an attitude deducing algorithm which is based on Kalman filter are present. In the simulation of the algorithm, the algorithm behaves excellent in low speed movement and low gravity environment in the surface of the moon. At the end of the chapter, the output of HRG is used in the simulation. The result indicates that the eular angle error is lower after FLP filtering if proper step size is chosen.