Kalman Filtering in the Presence of State Space Equality Constraints

We discuss two separate techniques for Kalman filtering in the presence of state space equality constraints. We discuss these methods and then prove that despite the lack of similarity in their formulations, under certain conditions, the two methods result in mathematically equivalent constrained estimate structures. We then discuss numerical stability for implementations. We conclude that the potential benefits of using equality constraints in Kalman filtering often outweigh the computational costs, and as such, equality constraints, when present, should be enforced by way of one of these two methods.