Self-tuning measurement fusion Kalman predictors and their convergence analysis |
| |
Authors: | ChenJian Ran |
| |
Affiliation: | Department of Automation , Heilongjiang University , Harbin, 150080, China |
| |
Abstract: | For multisensor systems with unknown parameters and noise variances, three self-tuning measurement fusion Kalman predictors based on the information matrix equation are presented by substituting the online estimators of unknown parameters and noise variances into the optimal measurement fusion steady-state Kalman predictors. By the dynamic variance error system analysis method, the convergence of the self-tuning information matrix equation is proved. Further, it is proved by the dynamic error system analysis method that the proposed self-tuning measurement fusion Kalman predictors converge to the optimal measurement fusion steady-state Kalman predictors in a realisation, so they have asymptotical global optimality. Compared with the centralised measurement fusion Kalman predictors based on the Riccati equation, they can significantly reduce the computational burden. A simulation example applied to signal processing shows their effectiveness. |
| |
Keywords: | multisensor information fusion measurement fusion parameter estimation self-tuning Kalman predictor information matrix equation convergence asymptotic global optimality dynamic variance error system analysis method |
|
|