|
H-infinity control strategy plays an essential role in modern
robust theory due to its capability of rejecting various
uncertainties, disturbances and noises. In this paper, a
variational approach is used to formulate a general state space
solution to multi-input multi-output (MIMO) H-infinity optimal
control problem, without the limitation placed by orthogonality
assumptions. The optimal H-infinity controller gain and observer
gain matrices are then obtained by solving two Riccati equations
via a recursive searching procedure. To achieve a prespecified
performance, a loop shaping scheme is formulated in this
research by adding three weighting functions to tune the
sensitivity, complementary sensitivity and power transfer
matrices to desire shapes. A H-infinity/LTR design procedure is
also formulated in this paper. The open loop transfer function
in and output feedback case is recovered to a chosen Target
Feedback Loop by adjusting the H-infinity dynamic compensator.
designed H-infinity/LTR methodology proposed in this paper
allows a prechosen closed-loop performance via a loop-transfer-
recovery (LTR) procedure so thattrade-off problems of
compensator robustness and the achieved disturbancerejection can
readily be solved. The design procedure discribed in this
paper is used to design automatic steering and diving control
system of a unmanned underwater vehicle. Using H-infinity/LTR
methodology, the tracking ability, uncertainty tolerance
andnoise attenuation can be achieved systematically.
|