LQ Control under Security Constraints

Abstract

In this paper, a LQ optimal control problem in considered under several security constraints. The security constraints depend on the information gathered by an adversary making partial noisy measurements of the state. Control sequences are designed to drive the system to a specified terminal state while minimizing a quadratic cost and satisfying bounds on the information available to the adversary. The resulting optimization problems are shown to be convex. Duality techniques are incorporated to compute the optimal solutions. For a terminal state with a continuous distribution, the optimal solution is shown to be affine. © Springer International Publishing Switzerland 2013.