Multisensor Suboptimal Fusion Student's $t$ Filter

A multisensor fusion Student's t filter is proposed for time-series recursive estimation in the presence of heavy-tailed process and measurement noises. It extends the single-sensor Student's t Kalman filter to the multisensor setup based on the suboptimal arithmetic average (AA) fusion approach which is driven from information-theoretic density fusion optimization and able to deal with unknown correlation among sensors. To ensure computationally efficient, closed-form t density recursion, moment matching approximation has been used for averaging the t densities aggregated from different sensors. Based on the same framework, we also extend the covariance intersection (CI) approach fort density fusion. Simulation demonstrates the strength of the proposed multisensor AA fusion-based t filter in dealing with outliers as compared with the classic Gaussian estimator, and the advantage of the AA fusion in comparison with the CI approach and the augmented measurement fusion.

Automated summary

A multisensor fusion Student's t filter is proposed for time-series recursive estimation in the presence of heavy-tailed process and measurement noises. The filter extends the single-sensor Student's t Kalman filter to the multisensor setup using the suboptimal arithmetic average (AA) fusion approach, which can handle unknown correlation among sensors. Simulation results show that the proposed multisensor AA fusion-based t filter is robust against outliers compared to the classic Gaussian estimator, and it outperforms the CI approach and augmented measurement fusion.