Research On Application Of Square Root Unscented Kalman Filter In Atomic Gravimeter

Gravitational acceleration is actually not a constant value,and the change of it is extremely complicated.Therefore,it can reflect a lot of information about the earth.The precise measurement of gravity plays an important role in many fields.The atomic gravimeter is a new generation of the main instrument for the measurement of gravity,and it can carry out long-term relative gravity observation and absolute gravity measurements.To obtain more accurate absolute gravitational acceleration is the main research work of this thesis.The research work of this thesis mainly includes:(1)Analysis of the process and principle of measuring the gravitational acceleration by the atomic gravimeter.In the atomic gravimeter,the Raman laser frequency needs to be changed in order to compensate for the Doppler shift caused by gravity,and the scanning rate of Raman laser frequency is defined as the chirp rate.The atomic interference fringe can be obtained by changing the chirp rate.The least squares method is usually used to fit the interference fringe to obtain the chirp rate.And then the value of the gravitational acceleration can be calculated according to the relationship between the chirp rate and the absolute gravitational acceleration.Therefore,the accurate estimation of chirp rate is the key issue to obtain accurate gravitational acceleration.(2)Analysis of least squares method and square root unscented Kalman filter.The least squares method does not consider the statistical characteristics of the measurement signal,and it only guarantees that the sum of the squared measurement error is minimized.And only under the condition that there is no system state error and the measurement error conforms to the normal distribution,a favorable estimated result can be obtained by the LS method.While,the Kalman filter makes full use of the statistical characteristics of the measurement signal.And it also considers the system state error and has no limitation on the error distribution.In addition,it can minimize the variance of the measurement error and the estimation error of the estimated parameter.The square root unscented Kalman filter is applicable tononlinear systems and can ensure the positive definiteness and symmetry of the matrix in the filtering process.Based on these analysis and comparison,this thesis proposes to apply the square root unscented Kalman filter to the gravity value estimation in the atomic gravimeter.(3)Analysis of the algorithm process and application of the square root unscented Kalman filter.This thesis describes the calculation process of the square root unscented Kalman filter and the adaptive noise estimation algorithm for real-time correction of measurement noise in detail.The simulated data and experimental data are processed by combining these two algorithms.And then the analysis and comparison of the results demonstrate the feasibility and effectiveness of the square root unscented Kalman filter in the estimation of gravity value.And the square root unscented Kalman filter can improve the accuracy of the gravity value.