Denoising the signal of the quartz flexural gravitational sensor by adaptive kalman filtering

JSD-I/A quartz flexural gravitational sensor (JQFGS) is an important component of the underwater navigation gravimeter, and the method of denoising the output signal of JQFGS by time series analysis and adaptive kalman filtering is researched. Firstly, the time-series models of JQFGS´s signal are analyzed, and AR(3), AR(4) and AR(5) models are used to approximate the output system equation of JQFGS. Secondly, Sage-Husa adaptive kalman filter is applied to denoise the output signal of JQFGS. The mean square deviations of JQFGS´s signal before denoising experiments is 2.352537×10^-9V and become 1.916178×10^-9V, 1.902123×10^-9V and 1.8891415×10^-9V after denoising experiments. The experiments results indicate that the method modeling the output signal of JQFGS by using time-series analysis, approximating the output system equation of JQFGS by using time-series model and denoising the signal of JQFGS by using Sage-Husa adaptive kalman filter is feasible.