1: \begin{abstract}                          % Abstract of not more than 200 words.

2: This paper addresses state estimation of linear systems 

3: with special attention on unknown process and measurement noise covariances, % with unknown noise statistics. 

4: aiming to enhance estimation accuracy while preserving the stability guarantee of the Kalman filter. 

5: To this end, 

6: the full information estimation problem over a finite interval is firstly  addressed. Then,

7: a novel adaptive

8: \textit{variational Bayesian} (VB) \textit{moving horizon estimation} (MHE) method is proposed, 

9: exploiting VB inference, MHE and Monte Carlo integration with importance sampling

10: for joint estimation of the unknown process and measurement noise covariances,

11: along with the state trajectory over a moving window of fixed length.

12: %Then, a bounded adaptive variational Bayes moving horizon (VB MHE) filter is developed

13: %by constructing the suppose that there are some prior bounds on the noise covariances. 

14: %Monte Carlo integration with improtance sampling is performed to approximate the expected inverse noise covariances 

15: %considering the problem on computing the covariance integrals in closed form.

16: Further, it is proved that the proposed adaptive VB MHE filter ensures mean-square boundedness 

17: of the estimation error 

18: with any number of importance samples and VB iterations, as well as for any window length.

19: Finally, simulation results on a target tracking example 

20: demonstrate the effectiveness of the VB MHE filter 

21: with enhanced estimation accuracy and convergence properties

22: compared to the conventional non-adaptive Kalman filter and other existing adaptive filters.

23: %trading off estimation performance versus sliding window length and stability.

24: \end{abstract}

25: