Many unmanned aerial vehicles (UAVs) can remain aerodynamically flyable after sustaining structural or control surface damage, yet insufficient robustness in conventional autopilots often leads to mission failure. This paper proposes a robust adaptive sliding mode controller (RASMC) for fixed-wing UAVs subject to aerodynamic coefficient perturbations and partial loss of control surface effectiveness. A damage-aware flight dynamics model is developed to systematically analyze the impact of such impairments on the closed-loop behavior. The RASMC is designed to ensure reliable tracking and stabilization, while a gain adaptation law maintains low control effort under nominal conditions and increases the gains as needed in the presence of aerodynamic damage. Lyapunov-based stability guarantees are derived, and assumptions on admissible uncertainty bounds are formulated to characterize the limits within which closed-loop stability and performance can be ensured. The proposed controller is implemented within an existing UAV autopilot framework, where outer-loop guidance and speed control modules provide reference commands to the RASMC for attitude stabilization. Simulations demonstrate that, despite significant aerodynamic damage and control surface degradation, all closed-loop states remain stable with bounded tracking errors.
disturbances, or structural damage. In [1], sliding-mode control (SMC) in combination with extended state observer is applied for robust attitude stabilization of fixed-wing UAVs. However, the approach does not explicitly account for aerodynamic uncertainties and is limited in the sense that it cannot handle degradation of control-surface effectiveness. Incremental dynamics-based SMC approaches are proposed by [2], [3] with consideration of actuator faults, including partial loss of control effectiveness, as well as structural damage. The stability analysis relies on lumped uncertainty bounds for the control effectiveness matrix, which is overly conservative and the proofs apply incremental control inputs within a time-continuous Lyapunov framework, which introduces technical challenges. In addition, the approaches use static gains and can not adapt to the amount of damage being present. The work of [4] employs a SMC-based backstepping approach for robust fixed-wing UAV control. It is capable of handling matched disturbances but assumes perfect knowledge of the control-surface effectiveness matrix. The authors in [5] design a SMC approach based on a linearized aircraft model that incorporates modeled damage. Although robustness is achieved, the control law requires prior knowledge of the structure of the damage-related input matrix, in contrast to most existing approaches, which assume only bounds on the uncertainties. In [6], robust control with respect to matched uncertainties is considered. Nonlinear dynamic inversion (NDI) in combination with a disturbance observer is applied for the rejection of the uncertainties.
Adaptive control methods aim to adjust the control laws in real-time based on changes in the system dynamics, such as those caused by damage [16]. These methods are particularly useful when the system is subject to unknown or varying conditions. In [7], a neural-network-based adaptive scheme is proposed to compensate for matched uncertainties in a fixed-wing UAV experiencing partial wing loss. The method is based on NDI and therefore assumes perfect knowledge of the control effectiveness matrix. The approach of [8] employs model reference adaptive control (MRAC) to compensate for matched uncertainties in the UAV inner-loop dynamics and was successfully validated in flight tests with a 25 % rightwing loss. The controller relies on inversion of a nominal model and assumes purely additive uncertainties. In [9], a control strategy for fixed-wing aircraft with asymmetric wing damage is proposed. Based on linearized dynamics a trimming algorithm is combined with an adaptive control allocation for an attitude PID controller. However, the PID gains themselves are not adapted in response to the damage. The authors in [10] address robustness against shifts in the center of gravity and perturbations in aerodynamic coeffi-cients for fixed-wing aircraft. Based on a linearized model, the approach combines NDI with a PI controller and an adaptive neural-network term. The adaptive component is designed to compensate only for matched uncertainties, while NDI assumes accurate knowledge of the control effectiveness matrix. The work of [11], [12] introduces a damage model based on perturbations and linearizes the aircraft dynamics around a nominal operating condition. A MRAC scheme is developed to compensate for changes in the system dynamics but relies on a linear approximation of the true nonlinear system.
While previous works have addressed robust or adaptive control for damaged UAVs, most assume perfect knowledge of the control effectiveness matrix. In addition, MRAC approaches rely on linearized dynamics, whereas most existing SMC methods can handle the full nonlinear dynamics but are not adaptive, requiring unnecessarily high gains.
In this paper, we present a robust adaptive sliding mode controller (RASMC) for nonlinear UAV dynamics under aerodynamic perturbations and degraded control surface effectiveness. We formulate an adaptation law for the controller gains, allowing them to remain low under nominal conditions and increase in the presence of damage. Lyapunov-based stability proofs are provided that guarantee the convergence of tracking error, offering formal assurances of stability even in presence of model mismatch. Additionally, we formulate conditions for the bounds of uncertainties under which closed-loop stability can be achieved, providing a clear framework for understanding the limits within which the proposed method remains stable. Finally, we demonstrate the effectiveness of the proposed method by embedding the RASMC into a full autopilot and simulating the UAV dynamics under damage.
The considered translational and rotational dynamics of the fixed-wing UAV are expressed in the body-fixed frame according to
with the states being: the body-fixed velocities V = u v w T ∈ R 3 , the rates around the axis of the bodyfixed frame ω = p q r T ∈ R 3 and the Euler angles
The inertia tensor is gi
This content is AI-processed based on open access ArXiv data.