Algebraic Stable Inversion of Non-Minimum-Phase Multivariable Control Systems
Stable inversion seeks a bounded control input that makes a system's output reproduce a prescribed trajectory. Casting the problem as a rational matrix equation and using the Smith normal form of the plant's polynomial matrix gives necessary and sufficient conditions for a bounded exact inverse even when the system is non-minimum-phase, non-square, or singular. When those conditions fail, an H-infinity-based approximation achieves near-perfect causal tracking without preview, and a dual-feedforward structure keeps the residual error bounded under uncertainty.
Algebraic Control: Complete Stable Inversion with Necessary and Sufficient Conditions
Stable inversion asks for a bounded control input that drives a linear time-invariant system's output to reproduce a prescribed trajectory. Working in an algebraic model-matching setting, the paper w…