Fault-Tolerant Control Based on Dynamic Regressor Extension and Mixing and Adaptive Allocation for Fixed-Wing Aircraft With Asymmetric Damage

Maintaining effective control of asymmetrically damaged aircraft is crucial for preserving the lives of the crew and mitigating property losses. However, the abrupt and substantial changes in aircraft parameters resulting from damage pose significant challenges to control. In this paper, a fault-tolerant control structure is proposed. Initially, the damaged aircraft model is reorganized and expressed in the form of an equivalent control component associated with the actuator plus a nonlinear function. Subsequently, an extended state observer is used to estimate the nonlinear function and derive a form suitable for parameter identification. A method based on improved dynamic regressor extension and mixing is then introduced to estimate the efficiency matrix of equivalent control. By incorporating a second filter and a smooth time varying forgetting factor, the persistent excitation condition is relaxed, rendering it suitable for parameter estimation in damaged aircraft scenarios. Finally, an adaptive control allocation algorithm is developed based on the estimated equivalent control efficiency matrix, with auxiliary adaptive variables used to effectively allocate virtual commands, even in cases of estimation error. To validate the efficacy of the proposed algorithm, we develop a fixed-wing aircraft with 40% damage to its right wing and conduct wind tunnel tests to obtain aerodynamic data before and after the damage occurs. Subsequently, numerical simulations and hardware-in the-loop experiments are performed. The results demonstrate that the proposed algorithm is capable of rapidly stabilizing and controlling the damaged aircraft, even when parameters are unknown, thus confirming its effectiveness and real-time performance.