Estimating gravity acceleration from static atomic gravimeter by Kalman filtering

We present the construction of the two-state model of the atomic gravimeter and the associated Kalman recursion to estimate gravity acceleration from atomic gravimeter. We find the Kalman estimator greatly improve the precision of estimates in short term by removing the white phase noise. The residual noise of the estimates follows 0.13 muGal/\sqrt{s} for up to more than 100 s and highlights a precision of 0.34 muGal at the measuring time of a single sample, even with no seismometer correction.