Reconfigurable LQR Framework for Optimality-Preserving Fault-Tolerant Control

When actuators lose control effectiveness due tomechanical damage or reduced motor efficiency, rotary-wingunmanned aerial vehicles (UAVs) become more prone toaccidents. Fault-tolerant control strategies for addressingsuch faults are broadly classified into Passive Fault-TolerantControl Systems (PFTCS) and Active Fault-Tolerant ControlSystems (AFTCS). While PFTCS are designed to handlepredefined fault scenarios, AFTCS actively estimate faultsonline and reconfigure the control system to recover desiredperformance. Linear Quadratic Regulator (LQR) controlis widely employed within AFTCS frameworks. However,experimental evidence reported in the literature indicatesthat, following actuator loss of effectiveness, reconfiguringthe system model using estimated fault parameters alone isinsufficient for restoring pre-fault performance. This limitationarises because the LQR control-weighting matrix R, designedfor the healthy system, no longer represents the true costof actuator usage under fault conditions. Although severalstudies have proposed reconfiguring the R-matrix to improvefault accommodation, these approaches do not guaranteepreservation of the pre-fault LQR optimality. In this work,a systematic method for reconfiguring the LQR R-matrixthat preserves pre-fault optimality is proposed. In addition,a reconfigurable feedforward control scheme is introduced toenhance reference-tracking accuracy under actuator faults.The proposed approach is validated on a two-degree-offreedom helicopter platform through both simulation andhardware-in-the-loop experiments, demonstrating satisfactoryfault-tolerant performance.