Output Complementary Sensitivity Shaping for MIMO Systems Without Multiple Experiments

Data-driven controller tuning methods have advantage that there is no modeling error and a designer can choice any controller structure. One of these methods, the noniterative correlation-based tuning (NCbT), needs multiple data acquisitions for MIMO systems. This paper proposes to reduce experiment number by regarding input complementary sensitivity function as an cost function and indirectly shaping output complementary sensitivity function by imposing a constraint for a condition number.