On the solution of a minimax terminal state estimation problem-A Hilbert space approach

The solution to a continuous-time LQG minimax terminal state estimation problem is obtained using a Hilbert space approach. The minimax decision rule is determined via the method of least favorable prior distributions. The minimax estimator is shown to be a linear functional of the observed data, and the least favorable prior distribution is shown to be a degenerate Gaussian stochastic process.