Decoupling approximation design using the peak to peak gain

Linear system design for accurate decoupling approximation is examined using the peak to peak gain of the error system. The design problem consists in finding values of system parameters to ensure that this gain is small. For this purpose a computationally inexpensive upper bound on the peak to peak gain, namely the star norm, is minimized using a stochastic method. Examples of the methodology's application to tensegrity structures design are presented. Connections between the accuracy of the approximation, the damping matrix, and the natural frequencies of the system are examined, as well as decoupling in the context of open and closed loop control.

► Decoupling for the large class of finite peak signals is important in practice. ► Decoupling uses the star norm which is computationally inexpensive. ► Stochastic optimization algorithm works well even for complex/large systems. ► Frequencies separation or diagonal dominance not necessary for accurate decoupling. ► Accurate decoupling is promising for open and closed loop control.