H∞ Optimal Controller Design With Closed-Loop Positive Real Constraints

Positive real constraints on the closed-loop of linear systems guarantee stable interaction with arbitrary passive environments. Two such methods of H-infinity optimal controller synthesis subject to a positive real constraint are presented and demonstrated on numerical examples. The first approach is based on an established multi-objective optimal control framework using linear matrix inequalities and is shown to be overly restrictive and ultimately infeasible. The second method employs a sector transformation to substitute the positive real constraint with an equivalent H-infinity constraint. In two examples, this method is shown to be more reliable and displays little change in the achieved H-infinity norm compared to the unconstrained design, making it a promising tool for passivity-based controller design.