Gain-Constrained Kalman Filtering for Linear and Nonlinear Systems
Bruno O. S. Teixeira, J. Chandrasekar, Harish J. Palanthandalam‐Madapusi, Leonardo A. B. Tôrres, Luís A. Aguirre, Dennis S. Bernstein · IEEE Transactions on Signal Processing · 2008
This paper considers the state-estimation problem with a constraint on the data-injection gain. Special cases of this problem include the enforcing of a linear equality constraint in the state vector, the enforcing of unbiased estimation for systems with unknown inputs, and simplification of the estimator structure for large-scale systems. Both the one-step gain-constrained Kalman predictor and the two-step gain-constrained Kalman filter are presented. The latter is extended to the nonlinear case, yielding the gain-constrained unscented Kalman filter. Two illustrative examples are presented.