The linear-exponential-quadratic-Gaussian control for discrete systems with application to reliable stabilization (Q1406082)

The paper studies a particular control algorithm for discrete-time, partially observable, linear stochastic systems. It uses an exponential performance criterion with the true states replaced by their estimates. As a consequence both the state and measurement noise covariances are considered in the computation of the control gain. This should provide good performance even in the presence of high noise environment. The algorithm is then applied in the context of the so-called reliable control design. An example of its application is finally examined.