Consensus variational Bayesian moving horizon estimation for distributed sensor networks with unknown noise covariances

Xiangxiang Dong,Giorgio Battistelli,Luigi Chisci,Yunze Cai
DOI: https://doi.org/10.1016/j.sigpro.2022.108571
IF: 4.729
2022-01-01
Signal Processing
Abstract:•Consensus filter is proposed for sensor networks with unknown noise statistics.•Variational Bayes inference is used to jointly estimate state and noise statistics.•Moving horizon estimation is adopted for possible bounds on the noise covariances.•Three consensus tasks on state and noise statistics are performed in parallel.•Existing consensus filters are generalized to parameter estimation over a window.
What problem does this paper attempt to address?