The ability to accurately determine the sensitivity of Riccati problems leads to the possibility of improving the condition of the Riccati equation, thereby improving the closed-loop response, by making changes in either the underlying system being controlled or by changing the weighing matrices for the state and input variables in the performance index. An immediate application of such a procedure involves the optimal positioning of sensors and actuators, particularly for structures with a large number of identical sensor/actuator mechanism, such as large space antennas. A related question is whether co-locating actuators and sensors always leads to the best system performance. Aspects of a method for determining optimal changes, in either the underlying system or the weighing matrices of the performance index, are discussed, as well as extensions to infinite dimensional distributed parameter systems.
