Sensor and Actuator Selection for Optimal Closed-Loop Performance in the Presence of Correlated Noise

A method is presented for the selection of sensors and actuators whose outputs may contain correlated noise. A generalization of the earlier results of Closed-Loop Input/Output Cost Analysis (CIOCA) is provided which utilizes decompositions of the closed-loop quadratic cost function into contributions from each stochastic noise input and each weighted output. It is shown that under certain circumstances the elimination of a correlated noise input may degrade performance. Application to a flexible space structure demonstrates the generalized method for sensor and actuator selection.