Pursuing a maneuvering target which uses a hidden Markov model for its control

For a partially observed linear pursuit-evasion problem in which the evader´s control is modeled as an unknown jump Markov process and a quadratic criterion involving the terminal miss, the optimal control is expressed as a linear function of the predicted miss. Determining the predicted miss involves estimating the evader´s terminal position from past system measurements. Expressions for unnormalized conditional probabilities are given which determine the conditional expectation of the evader´s terminal position.