Modal control of multi-input-multi-output systems

When the need for multi-input-multi-output control first arose, engineers already had at their disposal a solid set of frequency techniques suitable for single-input single-output control. Hence the temptation to use them:

• or directly, by creating as many feedback loops as there are input-output pairs (we'll assume throughout this section that m = r ), with each corrector designed to neglect the effect of interactions due to couplings z _ij ;

• or after designing an initial non-interaction corrector which, when cascaded to the system to be controlled, led to a diagonal or quasi-diagonal transfer matrix.

...