An improvement of the quantitative feedback theory for uncertain non-linear single-input-single-output (SISO) systems is obtained, by using a new procedure to convert the uncertain non-linear plant into an “equivalent” linear time-invariant plant set. The method is applied to the control of uncertain Van der Pol plant and Duffing plant. The results, compared with those obtained by following the classical approach, show the superiority of the new technique in reducing the “cost of feedback” in bandwidths of the compensation, for the same extent of plant uncertainty.
Read full abstract