A practical note on evaluating Kalman filter performance optimality and degradation

This paper presents useful remarks to the readers on the Kalman filter (KF) performance optimality, degradation, and some innovation related parameters. Guidelines for efficient approach for evaluation of the KF performance optimality and sensitivity analysis are presented. Performance degradation due to uncertainty in process and measurement noise statistics is discussed. Consistency check between the filter-calculated covariances versus actual mean square errors are provided, which can be used not only as a verification procedure for the filtering correctness, but also as a approach for making trade-off in designing a suitable Kalman filter. In addition to numerical algorithms, useful Matlab programs are accompanied where necessary to the readers for getting better insight in practical implementation. Exploration of the behaviour of some innovation based parameters useful in adaptive filter and system integrity designs, including covariance of innovation sequence, degree of mismatch (DOM), and degree of divergence (DOD), etc., is also involved.