System for improved receding-horizon adaptive and reconfigurable control

Abstract

A system for adaptive and reconfigurable control of aircraft and other complex objects and processes, herein referred to collectively as controlled “plants.” A modified form of sequential least-squares estimation is used to obtain near-real-time estimates of the values of plant parameters. The parameter estimates are used in a novel piecewise-linear optimal receding-horizon sampled-data controller in which internal controller parameters, known in the art as Riccati gains, are propagated from their current values using the most recent plant parameter estimates. This method of updating Riccati gains is a major departure from the prior art in receding-horizon control, where the Riccati gains (or equivalent) are re-initialized at each control update using terminal boundary conditions. The present invention provides improved robustness of control, lessened need for re-tuning control-effort penalties when operating conditions change, and significantly reduced computational burden. The control system of this disclosure is suitable for on-line adaptation to changing plant and control effector parameter values, including possible large and abrupt changes requiring controller reconfiguration for damage to or malfunction of the plant and/or control effectors.