Decentralized filtered dynamic inversion for uncertain minimum-phase systems

Decentralized filtered dynamic inversion is a control method for uncertain linear time-invariant systems that are minimum phase, have multi-input multi-output decentralized subsystems, and are potentially subject to unknown disturbances. This controller requires limited model information, specifically, knowledge of the relative degree and an estimate of the first nonzero Markov parameter for each local subsystem. Decentralized filtered dynamic inversion is effective for command following and rejection of unknown disturbances. We derive the decentralized filtered-dynamic-inversion controller and analyzes the closed-loop stability and performance (i.e., command-following error). We show that for sufficiently large choice of a single control parameter the closed-loop system is asymptotically stable, and the average power of the performance is arbitrarily small.