Review of SMC and FOSMC Strategies for Rotary Wing UAVs

Abstract

Unmanned Aerial Vehicles (UAVs) are widely used in fields such as autonomous missions, reconnaissance, surveillance, and various industrial applications. These vehicles can perform desired tasks without human intervention in challenging environmental conditions. However, UAV control can be difficult due to environmental factors, wind disturbances, and uncertainties in system parameters. Therefore, developing reliable control strategies for UAVs is a significant challenge for researchers and engineers. This study presents a comprehensive review of rotary-wing UAVs, focusing on quadcopter and helicopter systems. Approximately 77 studies were selected from the Web of Science (WOS) database and analyzed, with an emphasis on Sliding Mode Control (SMC) and Fractional-Order SMC (FOSMC) applications in these systems. The review addresses key topics such as degrees of freedom, proposed control methods, adjustment techniques, comparative methods, fractional-order definitions, simulation tools, and explanations. The literature analysis highlights current research trends by showing the performance advantages and limitations of SMC and FOSMC methods. Furthermore, future research directions and existing knowledge gaps are discussed in detail. This review was prepared to provide the control engineering community with a comprehensive understanding of SMC and FOSMC applications in rotary wing systems and to contribute to the development of innovative and effective control strategies.

1. Introduction

Unmanned Aerial Vehicles (UAVs), particularly rotary-wing platforms such as quadrotors and helicopters, have a wide range of applications in military, commercial, and civil sectors at present due to their vertical takeoff and landing capabilities, high maneuverability, and ability to operate in confined spaces [1]. These vehicles are effectively used in many critical areas such as environmental data collection, reconnaissance, search and rescue missions, and logistics [2]. However, parameter uncertainties, environmental effects, and dynamic variability encountered during the flight of these systems necessitate a safe and stable control mechanism [3]. Developing controller design is crucial for the safe and efficient operation of UAVs.

UAVs are significantly affected by external factors such as environmental conditions, wind changes, and uncertainties in system parameters, and traditional control methods may be insufficient to cope with these dynamics. Therefore, the development of nonlinear and robust control strategies has become an important area of research in the literature. Sliding Mode Control (SMC) is one of the control methods widely used in many nonlinear systems due to its robustness against disturbances and its ability to cope with uncertainties [4]. However, the chattering problem observed in classical SMC and the discontinuity in the control signal have necessitated the development of new control approaches.

For the purpose of reducing these issues, methods such as Adaptive SMC (ASMC) [5], Terminal SMC (TSMC) [6], Super Twisting SMC (STSMC) [7], Fuzzy SMC (FSMC) [8], and Fractional-Order SMC (FOSMC) [9] have been developed in the literature. In addition, more advanced control strategies have been proposed, such as Hybrid SMC [10], which is created by combining multiple approaches, and Hybrid FOSMC [11], which is developed by incorporating fractional-order definitions into hybrid structures. These advanced methods aim to improve control performance and overcome the limitations of classical SMC by combining different advantages.

Among these advanced control strategies, methods utilizing the notion of fractional order impart greater flexibility to the SMC structure and enable the incorporation of the system’s past information into the control design. In recent years, the addition of the fractional order notion to control theory has enabled the inclusion of memory effects in system dynamics through different derivative definitions such as Riemann–Liouville, Grünwald–Letnikov, Caputo, and Caputo–Fabrizio, thereby improving tracking and stability performance.

The purpose of this article is to conduct an in-depth examination of studies on the control of rotary-wing UAVs, specifically analyzing classical and fractional-order SMC-based control strategies used in quadrotor and helicopter systems.

An objective literature selection process was ensured by following a search strategy and selection procedure. The literature review was conducted using the Web of Science (WOS) database, covering the period from January 2015 to December 2025. The primary search string used was: (“Sliding Mode Control” OR “SMC”) AND (“Fractional Order” OR “FOSMC”) AND (“UAV” OR “Quadrotor” OR ‘Helicopter’ OR “2 DOF”) NOT (“Robot”). The search was then filtered to the years 2015–2025 and, as the research focused on article studies, publication types were limited to “article, early access” in the WOS database. In the next step, to focus on sliding mode control studies, the number of publications was limited by filtering the WOS database categories “Automation Control Systems”, “Engineering Aerospace”, “Engineering Multidisciplinary”, and “Mathematics Interdisciplinary Applications”. From the identified records, 77 articles were reviewed and selected for qualitative and quantitative synthesis to verify the performance of the methods and identify studies that present comparative results through simulation or experimental data.

The scope of the study is as follows:

• Fractional-order derivative definitions are explained.
• An examination of advanced methods such as Hybrid SMC, Adaptive SMC, and Terminal SMC for quadrotor systems and their fractional-order versions; and for helicopter systems, an examination of developed controller structures such as Adaptive SMC, Super Twisting SMC, Hybrid FOSMC, and Fuzzy FOSMC is presented.
• Research trends in the literature are presented in graphs.
• The strengths and weaknesses of existing studies are discussed, and recommendations for future research directions are presented.

The remainder of the article is organized as follows: The theoretical background for fractional computing and related derivative definitions are addressed in Section 2. Section 3 contains a comprehensive literature review and graphical trend analysis. Section 4 presents detailed dynamic models of the 6 degree of freedom (DOF) Quadrotor and 2 DOF Helicopter systems, which are the most frequently studied among rotary-wing UAVs in the literature. A general evaluation of the control strategies developed for Quadrotor and Helicopter systems is provided in Section 5. Finally, Section 6 concludes the paper with a discussion of the main findings and future research directions.

2. Preliminaries of Fractional Calculus

Fractional calculus emerged with the generalization of the definitions of integer-order derivatives and integrals to real or complex orders. The fractional operator ${\mathcal{D}}_t^{\alpha }$ of a function $f\left(t\right)$, generalized to order $\alpha$, is denoted as:

$${\mathcal{D}}^{\alpha }f\left(t\right)=\left\{\begin{array}{cc}{\displaystyle \frac{d^{\alpha }}{d{t}^{\alpha }}}f\left(t\right),\hfill & \alpha >0,\hfill \\ f\left(t\right),\hfill & \alpha =0,\hfill \\ {\displaystyle {\int }_0^{t}f\left(\tau \right)d\tau ,}\hfill & \alpha <0.\hfill \end{array}\right.$$

(1)

Defining the controller using a fractional operator allows not only the system’s current state but also its previous behaviour to be taken into account. This enables system responses to be adjusted more precisely by accounting for the memory effect, which is often overlooked by classical controllers. Significant advantages such as reduced transient regime durations, decreased overshoot rates, and suppression of chattering in the control signal ensure that fractional-order control strategies stand out in engineering applications.

The role of different fractional derivative definitions is also significant in the application of fractional calculus. Factors such as the physical interpretation of initial conditions, the properties of kernel functions, and the way the system’s memory behavior is represented have influenced the emergence of these definitions.

Table 1. Comparison of advantages and disadvantages according to fractional derivative definitions.

Fractional Derivative Definition	Math Expression	Advantages	Disadvantages
Riemann–Liouville (RL) [12]	$$D^{\alpha }f\left(t\right)={\displaystyle \frac{1}{\Gamma (n-\alpha )}}{\displaystyle \frac{d^n}{d{t}^n}}{\int }_a^t{(t-\tau )}^{n-\alpha -1}f\left(\tau \right)d\tau$$	- Strong definition in terms of mathematical proofs. - Widely used in theoretical analyses. - All memory history into account.	- The initial conditions cannot be physically defined. - May be sensitive to noise. - Because of the increasing computational load, it makes it difficult to apply to the control of physical systems.
Grünwald–Letnikov (GL) [12]	$$D^{\alpha }f\left(t\right)=\underset{h\to 0}{lim}{\displaystyle \frac{1}{h^{\alpha }}}\sum _{r=0}^{\lfloor t/h\rfloor }{(-1)}^r\left({\displaystyle \genfrac{}{}{0pt}{}{\alpha }{r}}\right)f(t-rh)$$	- Directly applied in discrete-time systems. - Applied in discrete sum form without requiring a continuous integral. - Suitable for numerical simulation applications. - Easy to implement in terms of numerical computation and programming. - Long-term memory feature like RL.	- Quite sensitive to noise. - The long-term memory effect makes it difficult to implement in real systems. - Weak stability in continuous systems.
Caputo (C) [12]	$$D^{\alpha }f\left(t\right)={\displaystyle \frac{1}{\Gamma (n-\alpha )}}{\int }_a^t{\displaystyle \frac{f^{\left(n\right)}\left(\tau \right)}{{(t-\tau )}^{\alpha -n+1}}}d\tau$$	- Initial conditions are easily defined in physical systems. - Has a strong memory effect. - Can be directly related to classical derivatives.	- The computational load is high. - Sensitive to noise. - May be slow in real-time applications.
Caputo–Fabrizio (CF) [13]	$$D^{\alpha }f\left(t\right)={\displaystyle \frac{M\left(\alpha \right)}{1-\alpha }}{\int }_a^{t}f\prime \left(\tau \right)e^{[-{\displaystyle \frac{\alpha (t-\tau )}{1-\alpha }}]}d\tau$$	- Provides short-term memory effect thanks to the exponential core. - Resistant to noise. - Numerically stable. - Suitable for real-time control applications.	- May respond slowly to dynamic changes. - Parameter settings are sensitive. - Performance may decrease in systems requiring long memory effects.

lists the commonly used definitions, including the Riemann–Liouville (RL), Grünwald–Letnikov (GL), Caputo, and Caputo–Fabrizio (CF) derivatives, along with their advantages and disadvantages. Here, $\Gamma \left(\right)$ is Euler’s Gamma function.

In the UAV control literature, the choice of fractional-order operators depends on whether the study is theoretical or application-oriented. The RL definition finds extensive use in theoretical analyses due to its mathematical rigor and its ability to fully account for the entire memory history. However, the inability to physically define initial conditions in the RL definition limits its direct use in physical systems. In contrast, the Caputo definition is more easily associated with physical systems, as it allows initial conditions to be expressed through directly measurable integer-order derivatives such as position and velocity. Furthermore, the GL definition offers a practical solution in numerical simulation and programming stages due to its discrete sum form, while the CF definition is evaluated in real-time applications requiring noise resistance thanks to its exponential kernel. Consequently, the choice of operator is not a random preference; it is based on the technical alignment between the mathematical advantages of the definition and the requirements of the targeted application platform.

3. Survey with Trend Analysis

In this section, based on graphical data derived from the literature review compiled in

Table 2. Summary of literature focused on SMC for Quadrotor.

Ref.	Year	DOF	Proposed Methods	Tuning/Manual	Compared with	E, S or Both?	Remarks
[14]	2015	6	SMC	Trial and error	$——$	Simulation	The study presents an SMC-based control architecture that maintains high tracking performance despite model uncertainties and measurement noise. Furthermore, the simplicity and implementability of the proposed design support its integration into mini UAV platforms.
[15]	2015	6	Immersion and Invariance-based Adaptive Control (for position) SMC (for attitude)	Adaptive online	$——$	Experimental and Simulation	This study proposes a hybrid control strategy that ensures high tracking performance under uncertainties and external disturbances by retaining the signum function in real-time experimental implementation, thereby demonstrating both theoretical stability and practical applicability.
[16]	2016	6	Adaptive RBF Neural Networks Control (for position) Integral SMC (for attitude)	Trial and error (for constant parameters) and Lyapunov-based Adaptive (online for RBFNN weights)	PD-Integral SMC, Backstepping-Nonlinear $H_{\infty }$	Simulation	This study presents a hybrid control strategy that ensures accurate trajectory tracking and high robustness under parametric uncertainties and external disturbances (including Gaussian noise). Comparative analyses demonstrate that the proposed method outperforms existing approaches in terms of tracking accuracy, fast convergence, and stable control signals.
[17]	2016	6	Backstepping SMC (for position) SMC (for attitude)	Trial and error	Standart Backstepping Control	Simulation	This study applies the proposed hybrid control method to a non-simplified quadrotor model in order to achieve position and attitude tracking performance under external disturbances, while an observer-based fault estimation module designed for the takeoff phase enhances system reliability.
[18]	2017	6	Global fast dynamic Terminal SMC	Analytical tuning via Hurwitz stability criterion	Terminal SMC and SMC in [19]	Simulation	The proposed control method exhibits smoother control signals and improved tracking performance compared to the reference controller in simulations conducted under external disturbances such as wind and air drag.
[20]	2017	6	Integral Backstepping SMC	Trial and error	PID, LQR, Backstepping, Integral-Backstepping	Simulation	The study integrates integral backstepping and SMC on a non-simplified quadrotor model for position and attitude tracking under external disturbances, and demonstrates improved performance in terms of tracking accuracy and control signal smoothness.
[21]	2018	6	Adaptive SMC	Adaptive online	Adaptive Integral Second-Order SMC in [22]	Simulation	The study proposes an adaptive sliding mode control approach incorporating PI sliding surfaces to ensure finite-time stability under parametric uncertainties, and demonstrates meaningful improvements in tracking convergence rate and control signal smoothness through comparative analysis.
[23]	2018	6	Robust Adaptive Second-Order SMC (for altitude and attidude)	Lyapunov-based online adaptation	Standart Adaptive SMC in [24], Super Twisting SMC in [25], Modified Super Twisting SMC in [25], Nonsingular Terminal SMC in [26]	Simulation	The adaptive second-order SMC method developed based on a PID sliding surface exhibits successful performance outcomes in terms of recovery time, oscillation level, and chattering effect under disturbances and model uncertainties; and delivers comparable or superior results compared to previously used methods in comparative analyses.
[27]	2019	6	Adaptive Backstepping Control (for position) Adaptive Backstepping Fast Terminal SMC (for attitude)	Lyapunov-based online adaptation	Integral Backstepping SMC in [28], Second-Order SMC in [29]	Simulation	The control architecture based on backstepping and fast terminal sliding mode techniques, structured through adaptive laws, has demonstrated stable performance under disturbances and uncertainties in terms of chattering reduction, fast convergence, and tracking accuracy, as confirmed by comparative analyses with classical and advanced methods.
[30]	2020	6	Robust Adaptive Nonsingular Fast Terminal SMC	Online Adaptive gain	Backstepping SMC in [17,31], Integral Backstepping SMC in [20], Feedback Linearization technique in [32]	Simulation	This study proposes an adaptive terminal SMC approach that effectively suppresses the chattering effect and ensures finite-time stability, providing high tracking accuracy and strong robustness for underactuated quadrotor systems subjected to external disturbances and model uncertainties.
[33]	2020	6	Super Twisting Second-Order SMC	Manual	SMC, Type 2 Fuzzy Logic-based controller	Simulation	This study proposes an effective control strategy based on a second-order sliding mode with super twisting algorithm, minimizing chattering while ensuring high-accuracy position and attitude tracking in a dual-loop quadrotor control architecture.
[34]	2021	3	Adaptive Super Twisting Nonsingular Terminal SMC	Adaptive online based on regulation strategy	Adaptive Nonsingular Terminal SMC, Nonsingular Fast Terminal SMC in [35]	Simulation	The proposed method combines the finite-time convergence and singularity avoidance features of the Adaptive Nonsingular Terminal SMC structure for attitude and altitude control, with the chattering reduction capability of the Adaptive Super Twisting SMC algorithm. This framework is applied to a quadrotor system subject to bounded disturbances, and the simulation results indicate that the method yields favorable outcomes in terms of control signal smoothness, tracking performance, and disturbance rejection when compared to other approaches.
[36]	2021	6	Neural Network-based Adaptive SMC	Neural Network + Backpropagation Adaptive online	Adaptive SMC in [24]	Simulation	In this study, a control structure incorporating a time-varying sliding surface, neural network-based online gain adaptation (via backpropagation rule), and disturbance observer integration achieves notable improvements in tracking accuracy and resilience against external disturbances, compared to the method used for comparison.
[37]	2021	6	Adaptive Fuzzy Terminal SMC	Online adaptive via Mamdani-type fuzzy logic	PID	Experimental and Simulation	The proposed method provides real-time adaptation capability, finite-time convergence, and robustness against model uncertainties and external disturbances, without requiring prior knowledge of system parameters. This framework has been applied to a quadrotor system subject to bounded disturbances; simulation and experimental data indicate that it yields more consistent tracking performance with lower error metrics in terms of control signal smoothness, tracking accuracy, and disturbance rejection compared to the referenced method.
[38]	2022	6	Neuro-Adaptive Fast Integral Terminal SMC	Feed-Forward Neural Network-based online adaptation	Second-Order SMC in [29], Integral SMC in [39]	Experimental and Simulation	The proposed control scheme integrates Terminal SMC, a robust exact differentiator, and a feed-forward neural network-based estimation structure to deliver a flexible strategy that operates online without requiring prior knowledge of the system model, exhibiting real-time learning capability and strong resilience against external uncertainties—thereby offering a noteworthy contribution to the literature.
[40]	2022	6	Adaptive RBF Neural Network SMC	RBFNN-based online adaptation	SMC in [41]	Simulation	The study integrates a classical SMC scheme with an RBF-based neural estimator to deliver a resilient and online adaptive control strategy that operates without requiring prior knowledge in the presence of varying payloads and model uncertainties.
[42]	2022	6	Integral Adaptive SMC	Adaptive online	Benchmark Integral Adaptive SMC	Simulation	The study presents a robust and effective control strategy that integrates an integral SMC approach with online adaptive gain tuning and a hyperbolic tangent-based chattering mitigation technique, ensuring reliable performance under variable payloads and model uncertainties without requiring prior system knowledge.
[43]	2022	6	Fuzzy SMC	Trial and error Genetic Algorithm	SMC	Simulation	This study presents a robust solution that integrates an optimized Fuzzy SMC structure to mitigate the drawbacks of SMC, thereby introducing an innovative control approach that enhances both tracking performance and robustness against external disturbances in quadrotor systems.
[44]	2023	6	Advanced Second-Order SMC	Manual	SMC, Second-Order SMC	Simulation	The study enhances the classical Second-Order SMC structure with an advanced reaching law and strict Lyapunov stability, and applies the resulting proposed algorithm to the Parrot Mambo quadrotor model. The proposed approach achieves superior simulation-based control performance in position and attitude tracking by providing lower tracking error and smoother control signals compared to the benchmarked methods.
[45]	2023	6	Adaptive Super Twisting Reaching Law SMC	Manual	SMC, Traditional Super Twisting Algorithm	Experimental and Simulation	The study integrates the proposed algorithm—based on an exponential adaptive law—into the classical SMC structure, and applies it to the Parrot Mambo quadrotor model. Compared to other benchmark methods, the proposed approach demonstrates superior control performance by yielding lower tracking error and smoother control inputs, as validated through both simulation and real-time hovering experiments.
[46]	2023	6	Terminal SMC	Manual	SMC in [47]	Simulation	The study integrates a terminal sliding surface, a quaternion-based compact structure, and a feedforward neural network compensator into the classical SMC framework, demonstrating superior control performance with lower tracking error and smoother control signals in comparative analyses.
[48]	2024	6	Adaptive SMC	Adaptive online	$——$-	Simulation	The study proposes an advanced SMC-based control structure that achieves stable tracking despite parameter uncertainties, input saturation, and external disturbances, by incorporating adaptive laws and a disturbance observer.
[49]	2024	6	Fuzzy Super Twisting SMC	Adaptive online	SMC, Fuzzy SMC, Fuzzy Super Twisting SMC	Simulation	The study proposes the integration of a fuzzy-based PID surface with a super twisting SMC structure to achieve superior tracking performance, offering lower tracking error and smoother control signals even in the presence of parameter uncertainties and external disturbances.
[50]	2024	6	Model-free Adaptive SMC	Model-free online adaptation	Genetic Algorithm, Particle Swarm Optimization-RBF, PID	Simulation	The study proposes a data-driven attitude control strategy based on a proposed structure, which operates solely on input–output data without relying on the quadrotor’s dynamic model, and combines chattering reduction with disturbance rejection to achieve high tracking accuracy and system stability.
[51]	2024	6	Fault Tolerant-Integral Terminal SMC	Trial and error	Integral Terminal SMC in [52]	Simulation	This study presents an integral terminal sliding mode-based fault-tolerant control design for quadcopter systems, which enhances flight safety by providing high robustness against actuator faults and achieving low tracking error.
[53]	2024	6	Predictor-based constrained fixed-time SMC	Predictive-optimal	Sliding Mode Predictive Control, Fixed-Time SMC	Experimental and Simulation	The study employs a predictor-based optimization approach incorporating fixed-time convergence and input-position constraints to achieve both high tracking performance and practical applicability in multi-UAV systems; furthermore, the algorithm’s real-time implementability is demonstrated through a Hardware-in-the-Loop test using Raspberry Pi 4.
[54]	2024	6	Antidisturbance SMC	Manual	Active Disturbance Rejection Control	Experimental and Simulation	This study proposes a novel reference model-based SMC approach, supported by a State Compensation Function Observer for state and disturbance estimation, to achieve accurate trajectory tracking under wind disturbances by compensating both the disturbance and its variation. The proposed method significantly improves control accuracy and system responsiveness.
[55]	2025	6	Super Twisting Algorithm Backstepping Adaptive Terminal SMC	Trial and error	Terminal SMC, Backstepping Terminal SMC, Adaptive Backstepping Fast Nonsingular Integral Terminal SMC in [56]	Simulation	This study demonstrates that the proposed approach, incorporating a super twisting algorithm for chattering suppression and an adaptive parameter structure to enhance control accuracy and system stability, provides significant performance advantages over existing methods in tracking tasks under rotor faults, model uncertainties, and external disturbances.
[57]	2025	6	Adaptive Neural Barrier-Based Terminal SMC	Adaptive online (for (RBF), Trial and error (for fixed parameters)	PID, SMC	Simulation	The proposed control approach, incorporating a BLF-based adaptive gain structure and online RBFNN compensation, achieves notable improvements in tracking accuracy and system stability under rotor faults and external disturbances, outperforming conventional methods.
[58]	2025	6	MLP-based Adaptive SMC	Adaptive online	SMC NN in [36], Conditioned Adaptive Barrier Function Integral Terminal SMC-Redfox in [59] Conditioned Adaptive Barrier Function Integral Terminal SMC-QPSO in [59]	Simulation	The control method developed with real-time MLP-based parameter adaptation demonstrates effective results in terms of recovery time, tracking accuracy, and energy efficiency, as evidenced by comparative scenarios involving both classical and optimization-based SMC approaches.

Ref.: References, DOF: Degree of freedom, SMC: Sliding Mode Control, UAV: Unmanned Aerial Vehicle; PI: Proportional-Integral, PD: Proportional-Derivative, PID: Proportional-Integral-Derivative, LQR: Linear Quadratic Regulator, NN: Neural Network, RBFNN: Radial Basis Function Neural Network, BLF: Barrier Lyapunov Function; MLP: Multi-Layer Perceptron, RBF: Radial Basis Function, QPSO: Quantum Particle Swarm Optimization E: Experimental, S: Simulation.

,

Table 3. Summary of literature focused on FOSMC for Quadrotor.

Ref.	Year	DOF	FO	Proposed Methods	Tuning/Manual	Compared with	E, S or Both?	Remarks
[60]	2017	4	RL	Fractional-Order SMC	Trial and error	$——$	Simulation	This study effectively demonstrates, through simulations, the potential of the fractional-order sliding mode control strategy to achieve fast convergence and stable performance in quadrotor systems under input saturation constraints.
[61]	2017	6	Caputo	Fractional-Order SMC	Trial and error	SMC	Simulation	The proposed method applied to the quadrotor adopts a double closed-loop structure comprising a fractional-order sliding mode controller in the inner loop and a PD-based controller in the outer loop, enabling simplification of complex dynamics while providing more accurate tracking and faster dynamic response compared to the referenced method.
[62]	2018	6	—	Fractional-Order Sliding Mode Observer-based Adaptive Fractional-Order Terminal SMC	Adaptive online	SMC	Simulation	The proposed method applied to the quadrotor, utilizing a two-layered control structure with proposed control in the inner loop and a PD controller in the outer loop, provides significantly improved tracking performance and chattering suppression compared to the benchmark method, even under partial actuator fault conditions.
[63]	2019	6	—	Adaptive Fractional-Order Sliding Mode-based Backstepping Controller	Lyapunov-based online adaptation	Sliding Mode-based Backstepping, Adaptive Sliding Mode-based Backstepping	Simulation	In this study, a control scheme enhanced with an adaptive correction coefficient and fractional-order sliding surfaces is proposed to improve the trajectory tracking performance of a quadrotor under mass and inertia uncertainties caused by varying payloads and wind disturbances. Simulation results indicate that, compared to benchmark control structures, the proposed method significantly reduces oscillations in position, attitude, and rotor speed, thereby increasing system robustness and improving tracking accuracy.
[64]	2019	6	RL	Fractional-Order Backstepping SMC	$——$	$——$	Simulation	The control approach presented in the study has been evaluated in simulation-based scenarios involving both translational and rotational dynamics under wind disturbances. The obtained findings indicate an effective tracking performance characterized by low tracking error, fast transient response, and high dynamic precision.
[65]	2019	6	RL	Nonsingular Terminal Fractional-Order SMC	Trial and Error	Classical terminal sliding surface ($S_1$), Linear sliding surface ($S_2$)	Experimental and Simulation	In this study, a FOSMC scheme incorporating a fractional-order, nonsingular terminal sliding surface is proposed for quadrotor UAVs under time-varying position constraints. A Lyapunov-based constrained controller is employed in the outer loop, while the inner loop uses the fractional-order strategy. The method is validated through simulations with sinusoidal disturbances—compared against $\left(S_1\right)$ and $\left(S_2\right)$ sliding surfaces—and real-time quad-directional experiments conducted on the Quanser QBall 2 platform. Results demonstrate that the FOSMC approach yields lower tracking error and faster convergence.
[66]	2019	6	Caputo	Fractional-Order SMC (for position) Fractional-Order Fast Terminal SMC (for attitude)	Trial and error	$——$	Simulation	In this study, a fractional-order finite-time hybrid control strategy is proposed for a quadrotor subject to uncertainties and external disturbances, aiming to track a complex flight trajectory. The proposed method is evaluated under nominal and $\pm 50\%$ parameter variations, where the use of a hyperbolic tangent function reduces the chattering effect, and simulation results demonstrate a rapid convergence behavior toward the sliding surfaces.
[67]	2020	6	RL	Adaptive Fractional-Order SMC	Trial and error	SMC	Simulation	Although the fundamental control strategy is common to both subsystems, the proposed control architecture regulates the position and attitude dynamics through distinct subcontrollers tailored to their respective physical characteristics. Under actuator faults and external disturbances, the tracking performance of the system is preserved; the finite-time convergence of all state variables has been established through Lyapunov-based analysis and corroborated by simulation results.
[68]	2020	6	RL	Fractional-Order SMC	Trial and error	Fractional-Order Fast Terminal SMC in [66]	Simulation	The proposed method is implemented in the same control form for both the position and attitude subsystems within a dual-loop control architecture; this structure enables the quadrotor to robustly track the reference trajectory under disturbances. Simulation results demonstrate that, compared to the referenced method, the proposed approach achieves lower tracking error, faster convergence, and smoother control signals.
[69]	2021	6	Caputo	Adaptive Fractional-Order Nonsingular Fast Terminal SMC	MATLAB Optimization Toolbox	Feedback Linearization in [32], Backstepping SMC in [17,31], Fractional-Order Backstepping SMC in [64], Nonsingular Fast Terminal SMC in [30]	Simulation	The proposed control structure, which simultaneously considers both translational and rotational dynamics of the mini quadrotor UAV, achieves high-precision tracking performance under random disturbances through the integration of fractional-order sliding surfaces and adaptive control laws with chatter-free control inputs. Simulation results demonstrate the superiority of the method over compared approaches in terms of shorter settling time, lower tracking error, and enhanced stability.
[70]	2021	6	Caputo	Improved Adaptive Fractional-Order Fast Integral Terminal Sliding Mode Control	Trial and error	Backstepping Sliding Mode Control in [17], Feedback Linearization in [32]	Simulation	The proposed method is independently applied to both the outer and inner control loops to ensure high robustness against external disturbances in quadrotor UAV systems. Simulation results demonstrate that, compared to existing control approaches, the method exhibits superior tracking performance with faster convergence, lower tracking error, reduced chattering effect, and smoother control signals.
[71]	2021	6	RL	Fractional-Order Global SMC	Simulink Optimization Toolbox	Fractional-Order Backstepping SMC in [64], İntegral Backstepping SMC in [20], Backstepping SMC in [17,31]	Simulation	For quadrotor systems operating under uncertainties and external disturbances, the proposed control approach provides superior control performance compared to benchmark methods by achieving fast convergence, high tracking accuracy, and reduced chattering effect.
[72]	2022	6	RL	Event-triggered Fractional-Order SMC	Lyapunov-based online adaptation (for gain parameters) Trial and error (for constant parameters)	Global Fast dynamic Terminal SMC in [18]	Simulation	The proposed controller guarantees system stability under external disturbances with low computational cost while avoiding Zeno behavior. Simulation results demonstrate that the system states converge to the sliding surface in finite time.
[73]	2022	6	RL	Fractional-Order Backstepping Fast Terminal SMC (for translational) Fractional-Order Fast Terminal SMC (for attitude)	Simulink Optimization Toolbox	Fractional-Order Backstepping SMC in [64], Integral Backstepping SMC in [20]	Simulation	This study is applied to a quadrotor system under external disturbances and parametric uncertainties, and proposes a fractional-order hybrid finite-time control structure. The conducted simulations demonstrate that the proposed method offers superior tracking accuracy and control performance compared to the referenced methods under the examined scenarios.
[74]	2023	6	Caputo	Fixed-Time Fractional-Order SMC	Trial and error	FOSMC, Backstepping SMC	Simulation	The proposed control structure offers a significant performance advantage over the compared methods in quadrotor trajectory tracking scenarios conducted under external disturbances, by ensuring fast convergence, low tracking error, and smooth control signals.
[75]	2023	6	Caputo	Fractional-Order SMC (for fractional model)	Trial and error	SMC (for integer model), FOSMC (for integer model), SMC (for fractional model)	Simulation	In this study, the rotational subsystem in all control configurations was governed using the classical SMC method. In contrast, the translational subsystem was explored under different configurations in terms of both modeling approach (integer-/fractional-order) and control technique (SMC/FOSMC). Simulation results demonstrate that the hybrid control approach based on the fractional-order translational model exhibited the highest performance among the compared methods in terms of tracking accuracy, chattering suppression, and response time.
[76]	2023	6	$——$	Fractional-Order SMC	Trial and error	$——$	Simulation	The direct coupling of the load dynamics into the quadrotor system enhances the applicability of the proposed control structure under practical conditions; simulation results indicate that the method is effective in suppressing chattering effects and improving the smoothness of control inputs.
[77]	2024	6	RL	Adaptive Barrier Function-based Backstepping Fractional-Order SMC	Trial and error	Adaptive Super Twisting SMC in [78]	Experimental and Simulation	This study proposes a backstepping-based fractional-order sliding mode control method integrated with an adaptive fractional-order barrier function for attitude and position tracking of quadrotor systems under external disturbances. The finite-time convergence capability of the proposed approach is validated through both MATLAB/Simulink simulations and real-time implementation on the Speedgoat platform.
[79]	2024	6	FOMCON toolbox	Fractional-Order SMC	Trial and error Genetic Algorithm	SMC	Simulation	The proposed method, tested under both GA-optimized and non-optimized configurations, demonstrated lower tracking error and more stable dynamic responses compared to the benchmark method in terms of tracking accuracy, transient behavior, and robustness against external disturbances.
[80]	2024	6	RL	Fractional-Order Fast Terminal Backstepping SMC (for inner loop) Fractional-Order Fast Terminal SMC (for outer loop)	$——$	$——$	Simulation	When the proposed dual-loop control structure is used together with a finite-time disturbance observer, simulation results show that the system can track the target trajectory within a 2% error margin under both constant and time-varying disturbances, and that disturbance effects are compensated in a short time.
[81]	2025	8	Caputo	Adaptive Finite-Time Fuzzy Fractional-Order Nonsingular Terminal SMC	FLS-based online adaptive parameter estimation	SMC, FOSMC in [82], Terminal SMC in [83]	Simulation	The fuzzy logic-based fractional-order adaptive sliding mode control strategy integrates antiswing architecture, adaptive tolerance against actuator faults, and slung load swing suppression within a unified framework, and demonstrates strong performance in terms of high tracking accuracy and stable control behavior in quadrotor UAV systems based on simulation results.
[84]	2025	6	Caputo	Fractional-Order SMC (for outer loop)	Trial and error	SMC, FOSMC, PID	Simulation	The propesed approach, which integrates fractional-order sliding mode control with active disturbance rejection in an inner loop structure, demonstrates notable improvement over existing methods in the literature by exhibiting high control accuracy and stability under severe disturbance conditions.
[85]	2025	6	Caputo	Fractional-Order SMC	$——$	Feedback Linearization	Experimental and Simulation	The proposed method enhances stability and robustness against external disturbances by explicitly incorporating uncertainty bounds into the system dynamics; thereby, it provides meaningful performance gains in quadrotor systems compared to the benchmark controller under real-world operating conditions.

Ref.: References, DOF: Degree of freedom, FO: Fractional-Order, SMC: Sliding Mode Control, FOSMC: Fractional-Order Sliding Mode Control, RL: Riemann–Liouville, UAV: Unmanned Aerial Vehicle; FLS: Fuzzy Logic System, PD: Proportional-Derivative, GA: Genetic Algorithm E: Experimental, S: Simulation.

,

Table 4. Summary of literature focused on SMC for Helicopter.

Ref.	Year	DOF	Proposed Methods	Tuning/Manual	Compared with	E, S or Both?	Remarks
[86]	2015	2	Optimal SMC	Quadratic Minimization technique	$——$	Simulation	The proposed approach, structured through quadratic optimization-based surface design and state-dependent feedback linearization, demonstrates high tracking performance and strong robustness against external disturbances on the TRMS.
[87]	2015	2	Integral SMC	Nonlinear least-squares Minimization	$——$	Experimental and Simulation	The experimentally validated multi-variable proposed approach, supported by extended Kalman Filter-based state and disturbance estimation, achieves high tracking performance and remarkable robustness against external disturbances on the TRAS platform.
[88]	2016	3	Adaptive Super Twisting SMC	Adaptive online	$——$	Experimental and Simulation	The experimentally validated proposed control approach, supported by a continuous differentiator, exhibits high tracking performance, remarkable robustness, and low control effort on a helicopter platform.
[89]	2017	3	Interval Type-2 Fuzzy SMC	Lyapunov-based analytical (for Observer gains), Literature-based Heuristic (for Fuzzy parameters)	Type-1 Fuzzy Logic Controller, Type-1 Fuzzy SMC, Interval Type-2 Fuzzy Logic Controller	Simulation	The proposed approach, supported by a nonlinear observer, demonstrates reduced chattering effects while exhibiting high tracking accuracy and strong robustness against model uncertainties in the helicopter system.
[90]	2018	2	Continuous Integral SMC-based CNF-STC	Lyapunov-based analytical (for CNF parameters), Literature-based Heuristic (for STC parameters)	Discontinuous Integral Sliding Mode-based CNF Control	Experimental and Simulation	The proposed structure, which integrates Composite Nonlinear Feedback (CNF) and Super Twisting Control (STC) methods within an Integral SMC framework, reduces chattering effects while providing high robustness against actuator saturation and fast, overshoot-free tracking performance in the helicopter system.
[91]	2018	2	Nonlinear Observer-based Fuzzy SMC	Manual	$——$	Experimental and Simulation	The proposed approach is structured to ensure stability in strongly coupled systems such as TRMS; designed to reduce control signal chattering, the method has been implemented in real-time on the TRMS platform using MATLAB Real-Time Toolbox and the Advantech PCI1711 data acquisition card, demonstrating its validity on the physical system.
[92]	2018	1	Terminal SMC	Approach based on Real-Time Data from 50 Experiments	$——$	Experimental	Distinguished by its ability to ensure stability without requiring the Brownian effect to vanish on the sliding surface, this original control approach has demonstrated effective performance on high-order stochastic systems despite its simple control law; experimental implementation through 50 trials on the actual system has verified that the system behavior remained reliably bounded under random perturbations.
[93]	2019	2	Adaptive Super Twisting SMC	Particle Swarm Optimization	Super Twisting SMC	Simulation	The proposed approach has been demonstrated through simulation-based results to provide successful tracking performance in multi-input aerodynamic systems, even under conditions with parameter uncertainties and external disturbances.
[94]	2019	2	SMC assisted by GPI Observer	Pole Placement via Characteristic Polynomial Assignment	Linear GPI Observer-based active disturbance rejection control	Experimental	The effectiveness of the proposed approach in providing successful tracking performance for multi-input aerodynamic systems has been experimentally validated, even under conditions involving parameter uncertainties and external disturbances.
[95]	2019	2	Super Twisting-based Nonlinear SMC	Quadratic performance index minimization (for linear surface gains)	PID	Experimental and Simulation	The controller, designed with a nonlinear sliding surface based on the variable damping ratio principle and employing the super twisting algorithm, achieves stable and low-chattering tracking performance in a multi-input aerodynamic system under simultaneous disturbances and structural uncertainties.
[96]	2021	2	Adaptive RBFNN Global Fast Dynamic Terminal SMC	Trial and error	Global Fast Dynamic Terminal SMC, SMC, PID	Experimental	The control approach, validated in real-time, demonstrates high tracking accuracy under external disturbances and wind effects, enabling stable and precise control of the TRMS without requiring prior knowledge of its dynamic model. Compared to other methods, it reveals distinct structural advantages in terms of disturbance robustness and ease of implementation.
[97]	2023	2	Optimal Second-Order SMC based on Metaheuristics and Artificial Neural Networks	Genetic Algorithm Deep Learning-Based Optimization	Second-Order Discrete SMC	Simulation	This study proposes a hybrid and innovative control approach that enhances tracking accuracy and system stability by introducing a novel sliding surface design based on the Sylvester equation, optimizing the LQR weighting matrices using GAs, and generalizing these solutions through an artificial neural network.
[98]	2023	2	Adaptive Neural Network SMC	Lyapunov-based online adaptation	PD, SMC	Experimental and Simulation	This study presents a hybrid sliding mode control approach that suppresses uncertainties through the integration of a RBF neural network and a saturation function, performs online parameter adaptation via an adaptive learning mechanism, and is validated through experimental implementation.
[99]	2023	6	Integral Backstepping SMC combined with Neural Network	Trial and error	Backstepping SMC, Integral Backstepping SMC without Neural Network	Simulation	The proposed control scheme applied to a medium-class unmanned autonomous helicopter exhibited robust tracking performance under input saturation, external disturbances, and system uncertainties; comparative simulation results demonstrated favorable tracking behavior with fast response time and limited oscillation.
[100]	2024	2	Adaptive Optimized Nonsingular Terminal Sliding Mode Super Twisting Control	Grey Wolf Optimizer, Whale Optimization Algorithm, Salp Swarm Algorithm, Ant Lion Optimizer	Optimized Nonsingular Terminal Sliding Mode Super Twisting Control	Simulation	The Nonsingular Terminal SMC-based hybrid control structure developed for the Quanser aerial simulator helicopter was designed as proposed control through the integration of metaheuristic optimization and adaptive super twisting techniques. The comparative simulation results revealed that among the employed metaheuristic algorithms, GWO provided the most effective optimization outcome. In the subsequent stage, comparative simulations supported by tests conducted in the ROS-Gazebo environment validated that the proposed structure is the most effective approach in terms of performance and robustness.
[101]	2024	2	Fuzzy logic enhanced Second-Order Sliding Mode Control	Multi-Objective Genetic Algorithm (for gain parameters), Fuzzy Logic (for sliding surface parameters)	SMC, Second-Order SMC	Experimental	The proposed control structure based on the Super Twisting Algorithm effectively suppresses chattering through fuzzy logic-based online parameter updating and demonstrates superior tracking performance compared to conventional methods in real-time experiments conducted on the TRMS.
[102]	2024	2	Variable Gain SMC	Trial and error	Traditional Twisting Algorithm-based SMC	Simulation	To evaluate the robustness of the proposed control structure for the TRMS, Gaussian white noise was incorporated into the model. By dynamically adjusting the gains, the control signal overestimation caused by model uncertainties is mitigated, and the control effort is minimized. The effectiveness of the controller was validated through simulations conducted in the MATLAB/Simulink environment.
[103]	2025	3	Observer-based Fixed-Time SMC	Genetic Algorithm, Nelder-Mead optimization, SSE algorithm	Continuous Nonsingular Terminal SMC in [104]	Experimental	The proposed approach was developed to achieve accurate trajectory tracking of the experimental helicopter platform and was integrated with a fixed-time extended state observer to enhance robustness against uncertainties and external disturbances. The effectiveness of the method was validated through comprehensive experimental tests and comparative analysis with an existing method in the literature.
[105]	2025	2	Sliding Mode Finite-time Disturbance Observer-Based Control	Gradient Descent algorithm	PID	Experimental and Simulation	In this study, a novel laboratory-scale TRMS system was developed; based on experimental data, a disturbance observer-based control structure with finite-time convergence was proposed, and the effectiveness of the proposed approach was thoroughly validated through both simulation and experimental scenarios.

Ref.: References, DOF: Degree of Freedom, SMC: Sliding Mode Control, TRMS: Twin Rotor Multiple-input and Multiple-output System, TRAS: Twin Rotor Aerodynamic System, LQR: Linear Quadratic Regulator, PD: Proportional-Derivative, PID: Proportional-Integral-Derivative, GPI: Generalized Proportional Integral, RBFNN: Radial Basis Function Neural Network, ROS: Robot Operating System, CNF: Composite Nonlinear Feedback, STC: Super Twisting Control, GA: Genetic Algorithm, RBF: Radial Basis Function, GWO: Grey Wolf Optimization, SSE: Sum of Squared Errors, E: Experimental, S: Simulation.

and

Table 5. Summary of literature focused on FOSMC for Helicopter.

Ref.	Year	DOF	FO	Proposed Methods	Tuning/Manual	Compared with	E, S or Both?	Remarks
[106]	2016	2	—	Type-2 Fuzzy Neural Networks Fractional-Order SMC with $P{D}^{\alpha }$	Lyapunov-based online adaptation	$P{D}^{\alpha }$, Type-2 Fuzzy Neural Networks with PD	Experimental and Simulation	The type-2 fuzzy controller structured with a FOSMC-based adaptive learning algorithm featuring a $P{D}^{\alpha }$ defined fractional sliding surface demonstrates superior tracking performance in the trajectory tracking task of a chaotic spacecraft; furthermore, its practical applicability is validated through real-time implementation on a helicopter using the Tustin approximation within a low-cost embedded system.
[107]	2019	2	GL	Fractional-Order SMC	Trial and error	SMC	Simulation	A fractional-order sliding surface for the TRMS is defined, and the control input is derived through system decoupling. Simulation results demonstrate that the chattering effect is significantly suppressed compared to the classical SMC.
[108]	2020	2	Caputo	Optimized Fractional-Order SMC	Genetic Algorithm	$——$	Experimental and Simulation	The proposed control structure was implemented on a Quanser AERO helicopter testbed; the fractional order of the sliding surface and the controller parameters were determined based on a cost function that minimizes chattering and tracking errors. Numerical and experimental results confirm the controller’s effectiveness in tracking the reference trajectories.
[109]	2023	2	—	Fractional-Order PID (for main rotor)+Fractional-Order Super Twisting SMC (for tail rotor)	Trial and error	PID, Fractional-Order PID, SMC, Super Twisting SMC, Fractional-Order Super Twisting SMC	Experimental and Simulation	In this study, for the first time, a hybrid control strategy tailored to the rotor dynamics of a TRMS system was developed by simultaneously applying two different fractional-order controllers within the same system, and this approach was shown to provide a significant and experimentally validated improvement in system performance compared to other methods.
[110]	2024	2	GL	Fractional-Order Fuzzy SMC using FRL	Fractional-Order Levenberg-Marquardt learning method	Fractional-Order Fuzzy SMC using IRL, Fractional-Order Fuzzy SMC using NE in [111]	Simulation	The proposed control structure adopts a tripartite architecture comprising the TSK-Fractional-Order Fuzzy SMC and TSK-Fractional-Order Fuzzy Equivalent Control actors, which respectively approximate the switching and equivalent control signals, along with the TSK-Fractional-Order Fuzzy Critic Network that estimates the value function. This controller was applied to a helicopter system, and the simulation results demonstrate its superiority over the compared methods in terms of error performance metrics (ISE and IAE).

Ref.: References, DOF: Degree of Freedom, FO: Fractional-Order, SMC: Sliding Mode Control, FOSMC: Fractional-Order Sliding Mode Control, GL: Grünwald–Letnikov, PD: Proportional-Derivative, PID: Proportional-Integral-Derivative, TRMS: Twin Rotor Multiple-input and Multiple-output System, FRL: Fractional-Order Reinforcement Learning, IRL: Integer-Order Reinforcement Learning, NE: Neural Estimator, TSK: Takagi–Sugeno–Kang, ISE: Integral of Squared Error, IAE: Integral of Absolute Error, E: Experimental, S: Simulation.

, the trends in modeling, application, and control strategies for Quadrotor and Helicopter systems from rotary wing aircraft, along with the fundamental conclusions drawn from these trends, are presented. The distribution of studies using SMC and FOSMC methods in the control of Quadrotor and Helicopter systems, according to the number of articles published between 2015 and 2025, is presented in Figure 1.

Looking at Table 2 and Table 3, the trend summaries for quadrotors are presented in detail in Figure 2, Figure 3 and Figure 4. According to Figure 2a, although Integer-Order approaches ($56.6\%$) are dominant in controller design, Fractional-Order methods are also preferred to a significant extent ($43.4\%$), indicating that researchers increasingly seek the additional tuning flexibility and memory characteristics provided by fractional calculus despite the continued prevalence of classical control frameworks. Looking at Figure 2b, it is observed that the vast majority of studies ($94.33\%$) were conducted on quadrotors with 6 degrees of freedom (DOF), suggesting that the full nonlinear 6 DOF model has been widely adopted as a standard reference framework for comprehensive performance evaluation and consistent comparison of control strategies. When the application methods are analysed in Figure 2c, it can be observed that the majority of the studies examined ($83.02\%$) were conducted in a simulation environment, indicating that simulation-based validation remains the dominant approach due to its cost-effectiveness, safety, and flexibility in testing various operating scenarios, although it may limit the assessment of real-world implementation challenges. In Figure 2d, the preferred fractional derivative definitions are shown; the Riemann–Liouville ($43.48\%$) and Caputo ($39.13\%$) definitions are the most commonly used, reflecting their strong theoretical foundations and widespread acceptance in control-oriented fractional calculus, which facilitates mathematical analysis and the comparison of different methods. The distribution of control methods presented in Figure 3 indicates that the literature is moving toward more modular architectures in order to overcome the classical limitations of SMC. The dominance of Hybrid SMC ($16.39\%$) and Adaptive SMC ($13.11\%$) reflects a tendency to combine multiple control layers (such as integral components or adaptive laws) to suppress the chattering effect while maintaining robustness under dynamic uncertainties. Furthermore, the preference for Terminal SMC ($8.2\%$) and Terminal FOSMC ($8.2\%$) demonstrates the increasing importance attributed to finite-time convergence in the design of high-precision UAV missions. This suggests that ensuring system states reach the equilibrium point within a predictable and bounded time interval has become a key design priority. Finally, examining the parameter tuning strategies in Figure 4, Trial and Error ($36.67\%$) emerges as the most commonly used approach. This high prevalence reflects the difficulty of establishing a direct analytical relationship between controller gains and the complex dynamics of UAVs, which has led researchers to rely on heuristic tuning methods. Accordingly, the growing interest in metaheuristic optimization techniques indicates that the literature is gradually shifting from manual tuning toward more systematic and automated procedures.

Looking at Table 4 and Table 5, the trend summaries for helicopters are presented in detail in Figure 5, Figure 6 and Figure 7, and it is observed that they exhibit different trends compared to quadrotors. According to Figure 5a, Integer-Order control approaches ($79.17\%$) are more widely preferred, while interest in Fractional-Order methods ($20.83\%$) remains limited. This distribution indicates that control performance under aerodynamic cross-couplings is ensured by integer-order controllers, which provide analytical stability and low computational overhead. On the other hand, the presence of fractional-order methods in the literature reflects academic interest in additional parametric degrees of freedom, which enable enhanced robustness and superior disturbance rejection in complex flight regimes. Analysis of Figure 5b reveals that a substantial majority of the studies focus on simplified 2 DOF systems ($79.17\%$). This concentration suggests that researchers prefer 2 DOF benchmarks to evaluate control performance under fundamental challenges such as cross-coupling and nonlinear effects. Consequently, these systems constitute a vital verification stage before the implementation of proposed control strategies on higher-dimensional dynamics. The application data in Figure 5c indicates that the Simulation + Experimental ($41.67\%$) combination is the most prevalent trend. This finding reveals that researchers do not limit their control strategies to simulation environments but instead support them with concrete data through testing under real flight conditions. Given the complex dynamics of helicopters and the presence of unmodeled disturbances, this dual-stage validation process verifies the practical feasibility, system stability, and robustness of the control performance. Figure 5d shows that the Grünwald–Letnikov definition and ’Not specified’ cases are equally predominant ($40\%$). However, the lack of clarity regarding the operator selection undermines the reproducibility of the methods, as different definitions (e.g., Caputo, RL, or GL) possess distinct memory and initialization properties. This ambiguity hinders the assessment of the specific operator’s impact on the system’s dynamic response and numerical accuracy, thereby limiting the overall scientific reliability of the proposed strategies. The distribution of control methods in Figure 6 indicates that Hybrid SMC ($20.37\%$), Adaptive SMC ($9.26\%$), and Super Twisting SMC ($9.26\%$) are the predominant variants. This trend confirms that the literature is focused on mitigating the chattering problem inherent in classical SMC and enhancing robustness against uncertain system parameters. Specifically, the preference for Super Twisting and Adaptive variants stems from the necessity to manage high-degree nonlinearities and external disturbances in helicopter dynamics without compromising control signal continuity. This suggests that the proposed methods provide high tracking precision while preserving actuator longevity. Figure 7, which presents the adjustment methods, indicates that Metaheuristic Optimization ($27.27\%$) is the most frequently employed approach in helicopter studies. This preference reflects the complexity and strong nonlinearities of helicopter dynamics, for which conventional analytical tuning becomes challenging, thereby encouraging the use of optimization-based strategies capable of systematically searching for near-optimal controller parameters under multiple performance constraints.

This trend analysis reveals that Quadrotor studies predominantly favour Hybrid SMC and 6 DOF modeling; Helicopter studies also prefer Hybrid SMC as a method, similar to Quadrotor studies, but place greater emphasis on 2 DOF modeling and experimental validation steps. In determining control parameters, there is a tendency towards the Trial and Error method in quadrotor studies and the Metaheuristic Optimization method in helicopter studies. These findings indicate that future research will focus on hybrid control methods that combine not only simulation studies but also experimental applicability.

4. Modeling of Rotary-Wing UAV Systems

This section presents the dynamic equations of the 6 DOF quadrotor and 2 DOF helicopter systems most widely utilized in the literature. Furthermore, it examines how the nonlinear nature of quadrotor and helicopter dynamics, parameter uncertainties, and the structural interactions between disturbance effects and control inputs play a critical role within the context of the impact of system dynamics on control design. These evaluations establish a fundamental framework for understanding the various control approaches analyzed in Section 5.

4.1. Quadrotor Model

As illustrated in Figure 8, 6 DOF quadrotor systems consist of four rotors positioned symmetrically on the airframe. The translational and rotational motions of the system are modeled using the Newton–Euler approach through the following equations [18,29]:

$$\left\{\begin{array}{c}\ddot{x}={\displaystyle \frac{1}{m}}(cos\varphi sin\theta cos\psi +sin\varphi sin\psi )u_1-{\displaystyle \frac{K_1\dot{x}}{m}}\hfill \\ \ddot{y}={\displaystyle \frac{1}{m}}(cos\varphi sin\theta sin\psi -sin\varphi cos\psi )u_1-{\displaystyle \frac{K_2\dot{y}}{m}}\hfill \\ \ddot{z}={\displaystyle \frac{1}{m}}(cos\varphi cos\theta )u_1-g-{\displaystyle \frac{K_3\dot{z}}{m}}\hfill \\ \ddot{\varphi }=\dot{\theta }\dot{\psi }{\displaystyle \frac{I_y-I_z}{I_x}}+{\displaystyle \frac{J_r}{I_x}}\dot{\theta }{\Omega }_r+{\displaystyle \frac{l}{I_x}}{u}_2-{\displaystyle \frac{K_{4}l}{I_x}}\dot{\varphi }\hfill \\ \ddot{\theta }=\dot{\psi }\dot{\varphi }{\displaystyle \frac{I_z-I_x}{I_y}}-{\displaystyle \frac{J_r}{I_y}}\dot{\varphi }{\Omega }_r+{\displaystyle \frac{l}{I_y}}{u}_3-{\displaystyle \frac{K_{5}l}{I_y}}\dot{\theta }\hfill \\ \ddot{\psi }=\dot{\varphi }\dot{\theta }{\displaystyle \frac{I_x-I_y}{I_z}}+{\displaystyle \frac{1}{I_z}}{u}_4-{\displaystyle \frac{K_6}{I_z}}\dot{\psi }\hfill \end{array}\right.$$

(2)

In these equations, $u_1$ represents the total thrust, m denotes the total mass of the quadrotor, and g is the acceleration due to gravity. The aerodynamic damping effects occurring during translational motions are incorporated into the model via the coefficients $K_1$, $K_2$, and $K_3$. While the angular positions $\varphi$, $\theta$, and $\psi$ define the roll, pitch, and yaw axes, respectively; $I_x$, $I_y$, and $I_z$ represent the moments of inertia of the airframe, and $J_r$ signifies the moment of inertia of each rotor. The control torques $u_2$, $u_3$, and $u_4$ are generated through the angular velocity differences of the rotors located at a distance l from the center of mass. Furthermore, the gyroscopic effects resulting from the rotational speeds of the rotors (${\Omega }_r$), along with the aerodynamic friction coefficients ($K_4,K_5,K_6$) associated with angular velocities, are among the primary disturbance factors affecting the system dynamics.

Figure 8. Quadrotor UAV.

Figure 8. Quadrotor UAV.

Impact of Quadrotor Dynamics on Control Design

The complex and nonlinear dynamic structure of the quadrotor system is a fundamental determinant in the selection and design of the control approaches detailed in Section 5. The fact that the system has only four independent control inputs against six degrees of freedom necessitates a strong coupling between position control in the x and y planes and the angular orientations. The trigonometric expressions observed in Equation (2) and this hierarchical structure cause the system to exhibit a highly nonlinear character.

The severe inter-axis interaction arising from Coriolis and centrifugal forces creates a complex control problem where traditional methods remain insufficient. To overcome these dynamic challenges, Hybrid SMC approaches prominent in the literature are widely preferred to suppress the nonlinear character at different operating points and to enhance robustness. Furthermore, uncertainties in the quadrotor’s mass distribution, moments of inertia, and aerodynamic damping coefficients directly affect controller performance. To compensate for these parametric variations and external disturbances online, Adaptive SMC approaches play a critical role in adapting the system to changing flight conditions.

Since the convergence of tracking errors to zero within a finite time is of vital importance in aviation applications, Terminal SMC approaches are considered an effective solution offering fast and sharp maneuverability. However, to reduce high-frequency chattering that can occur in classical SMC structures and lead to mechanical wear in propellers, Fractional-Order SMC (FOSMC) approaches are employed. The memory effect provided by fractional-order operators enhances the system’s robustness by offering a more flexible sliding surface against the gyroscopic moments ($J_r\dot{\theta }{\Omega }_r$) and inter-axis dynamic couplings in Equation (2).

Finally, Hybrid FOSMC approaches, which combine the advantageous features of different control methods to manage the complex dynamics of the quadrotor, stand out in the literature as an advanced control strategy providing both finite-time convergence and high robustness. Nevertheless, mass uncertainties and the delicate balance between the inner-outer loops render these advanced control structures highly sensitive to parameters. Although the statistical data presented in Section 3 indicates that the trial-and-error method is widely used, it is observed that this manual approach is insufficient to provide optimum results; this situation has directed the literature toward optimization methods. Thus, metaheuristic optimization plays a significant role in improving the flight performance of the quadrotor by ensuring the automatic tuning of hybrid and fractional-order parameters that are difficult to adjust manually through intelligent optimization loops.

4.2. Helicopter Model

The 2 DOF helicopter system presented in Figure 9 is a complex MIMO (Multiple-Input Multiple-Output) structure used to analyze motions around the pitch and yaw axes. Utilizing the Euler–Lagrange approach, the dynamics of the system are expressed by the following equations [93]:

$$\left\{\begin{array}{c}(J_p+m{l}_{cm}^2)\ddot{\theta }=K_{pp}{V}_p+K_{py}{V}_y-B_p\dot{\theta }-mg{l}_{cm}cos\theta -m{l}_{cm}^2{\dot{\psi }}^{2}sin\theta cos\theta \hfill \\ (J_y+m{l}_{cm}^2{cos}^2\theta )\ddot{\psi }=K_{yp}{V}_p+K_{yy}{V}_y-B_y\dot{\psi }+2m{l}_{cm}^2\dot{\theta }\dot{\psi }sin\theta cos\theta \hfill \end{array}\right.$$

(3)

In this model, $\theta$ represents the pitch angle and $\psi$ represents the yaw angle. $V_p$ and $V_y$ are the control inputs, specifically the motor voltages. $J_p$ and $J_y$ denote the moments of inertia for the respective axes, while $B_p$ and $B_y$ express the friction coefficients. The terms $K_{pp}$ and $K_{yy}$ determine the motor torque gains of the main axes, whereas $K_{py}$ and $K_{yp}$ are the gains that determine the intensity of the inter-axis cross-coupling. The parameter $l_{cm}$ represents the distance of the center of mass from the axis of rotation, determining the intensity of the gravitational torque. The term $mg{l}_{cm}cos\theta$ in the equation shows that the gravitational torque varies continuously depending on the pitch angle due to the position of the center of mass ($l_{cm}$).

Figure 9. Helicopter UAV.

Figure 9. Helicopter UAV.

Impact of Helicopter Dynamics on Control Design

The complex dynamic structure of the 2 DOF helicopter system plays a decisive role in determining the control approaches preferred in the literature. In particular, the terms $K_{py}{V}_y$ and $K_{yp}{V}_p$ in Equation (3) demonstrate that a control input applied to one axis directly affects the dynamics of the other, revealing the presence of inter-axis cross-coupling in the system. A review of the literature shows that methods such as Hybrid SMC and Fuzzy FOSMC stand out with their effective solution capabilities due to these strong dynamic interactions and the significant nonlinear character of the system.

Furthermore, Adaptive SMC approaches are widely discussed in the literature to eliminate the negative effects of uncertainties in friction coefficients ($B_p,B_y$) and unpredictable external disturbances on system performance. High-frequency chattering, which are frequently encountered in classical SMC applications and can lead to mechanical wear in physical system components, can be significantly reduced thanks to Super Twisting SMC algorithms developed to obtain a smoother and more stable control signal. In this respect, Super Twisting-based approaches are presented as an effective and practical solution in the literature.

Finally, changes in physical parameters such as the center of mass position ($l_{cm}$) and the high-order nonlinear structure of the system explain the interest in metaheuristic optimization methods emphasized in the statistical data presented in Section 3. These optimization techniques are observed to be evaluated as a critical design tool for the systematic tuning of numerous complex parameters found particularly in hybrid and fractional-order control structures.

5. Review, Categorization, and Comparative Analysis of SMC and FOSMC Approaches

This section provides a systematic review and categorization of SMC and FOSMC control approaches developed for various quadrotor and helicopter systems, based on the theoretical foundations presented in Section 2 and Section 3 and a comprehensive literature analysis. The primary objective of these control approaches is to achieve high accuracy, stability, and robust performance despite nonlinear system dynamics, model uncertainties, and external disturbances. As discussed in Section 3, these control architectures are categorized as subtypes of classical SMC (e.g., Super Twisting SMC, Terminal SMC, Adaptive SMC, etc.) and FOSMC-based approaches that combine these subtypes with fractional-order theory. Consistent with the method distributions presented in Figure 3 and Figure 6, detailed analyses of widely used control approaches, performance comparisons, and application examples from the literature are systematically presented below. Furthermore, the technical challenges and a comprehensive comparative analysis of all control strategies are provided in the final subsection to evaluate their practical feasibility in real-world applications.

5.1. Hybrid SMC Approaches for Quadrotor Systems

Hybrid SMC is an advanced control strategy created by combining multiple subtypes of the classic SMC method (Adaptive, Super Twisting, Terminal, Integral, Fuzzy, Neural or Backstepping, etc.) or different control approaches (PID, LQR, etc.) under a single controller structure. Furthermore, in hybrid SMC structures, applying different controllers to each loop in inner–outer loop or position–attitude control architectures also leads to the formation of a hybrid structure. For example, using Super Twisting SMC in the inner loop and Adaptive SMC in the outer loop SMC or a PID-based controller in the outer loop, the system transforms into a hybrid SMC structure combining different SMC or classical control approaches. This approach brings together the advantages of different control principles to enhance the system’s stability, robustness, and convergence speed. As Table 2 shows, the Hybrid SMC method, which uses different controllers together for inner-outer loop or position–attitude structures in the control of Quadrotor systems with six degrees of freedom, is generally compared with the different Hybrid methods given in the table with studies conducted in simulation or experimental–simulation environments. It has been demonstrated that the proposed methods exhibit superior performance in terms of tracking accuracy, fast convergence, and stable control characteristics [15,16,17,27]. In some studies where multiple subtypes of the classical SMC method were used together, it was noted that the Integral Backstepping SMC method exhibited better performance in terms of tracking accuracy and control signal smoothness when compared with PID, LQR, Backstepping, and Integral Backstepping control methods in a simulation environment [20]. The proposed Adaptive Super Twisting Nonsingular Terminal SMC method for controlling a three-degree-of-freedom Quadrotor system, when compared with the Adaptive Nonsingular Terminal SMC and Nonsingular Fast Terminal SMC methods in a simulation environment, has demonstrated better performance [34]. The Neuro-Adaptive Fast Integral Terminal SMC method has demonstrated real-time learning capability and strong robustness against external uncertainties when compared to Second-Order SMC and Integral SMC in experimental and simulation environments [38]. Other studies involving subtypes have also observed better results compared to the methods under consideration [23,30,33,36,37,40,42,45,49,50,51,55,57].

5.2. Adaptive SMC Approaches for Quadrotor Systems

Classic SMC ensures that the system can operate stably despite external influences and model uncertainties. However, due to fixed control gains, uncertainties in model parameters and external influences can create difficulties for system stability. Adaptive SMC, developed to overcome these difficulties, is a method that automatically adapts control gains based on the system state and error size. Some of the studies examined in Table 2 [21,48,58] applied classical Adaptive SMC to the control of Quadrotor systems with six degrees of freedom, and in these studies, the control parameters were determined using the Adaptive online method. Others [23,27,30,34,36,37,38,40,42,45,50,55,57] have used Adaptive SMC as a component of a hybrid SMC structure, and in these studies, control parameters have been determined using methods such as Adaptive online and Trial and error.

5.3. Terminal SMC Approaches for Quadrotor Systems

The Terminal SMC (TSMC) provides extremely fast stabilization performance by ensuring the system reaches the sliding surface within a finite time. This feature makes the method highly advantageous in conditions where the UAV’s rapid response is critical, such as sudden maneuvers or emergency situations. Examining Table 2, it is seen that in some studies [18,27,30,34,37,38,51,55,57], TSMC is used as a component of the hybrid structure, and the control parameters are determined using different methods depending on the components of the hybrid structure. In contrast, in [46], classical TSMC was applied instead of a hybrid structure, and the control parameters were determined manually. Findings in the literature indicate that the high convergence speed of TSMC’s increases the maneuverability of quadrotor systems and provides a more stable flight against external disturbances.

5.4. Terminal FOSMC Approaches for Quadrotor Systems

The Terminal FOSMC ensures that the system reaches the sliding surface within a finite time while incorporating past system dynamics into the control design using fractional definitions; thus enabling the controller to provide a more flexible and stable response. Examining Table 3, it can be seen that Terminal FOSMC is used as a component of the hybrid structure in studies [62,65,69,70,81], and that the control parameters are determined using different methods depending on the components of the hybrid structure. Furthermore, the Caputo definition is mostly preferred as the fractional derivative operator. In contrast, in studies conducted by [66,73,80], Terminal FOSMC is used in the inner loop [66] or in both the inner and outer loops [73,80], depending on the application architecture. Among these studies, the Caputo definition is seen to be preferred in [66], while the Riemann–Liouville definition is preferred in [73,80].

5.5. Classical FOSMC Approaches for Quadrotor Systems

Fractional-Order SMC (FOSMC) is an extended form of classical SMC that uses fractional definitions. This approach provides more flexible dynamic behavior and stability by incorporating the system’s past dynamics into the control design. Examining Table 3, it is seen that FOSMC is used in the studies [61,75,84,85] and in the outer loop of the study [66], and the Caputo definition is used as the fractional derivative operator. In studies [60,68,71], the Riemann–Liouville definition is preferred, and the Simulink Optimization Toolbox is used to determine the control parameters in study [71], while the trial and error method is used in the others. Study [76] does not specify the fractional derivative definition, and the parameters are determined using the trial and error method. In [79], both structures optimized with the genetic algorithm and unoptimized structures are tested, and therefore both the genetic algorithm and trial and error are used as parameter determination methods. Furthermore, in this study, the fractional derivative operator is defined theoretically, and the FOMCON Toolbox is preferred for numerical calculations.

5.6. Hybrid FOSMC Approaches for Quadrotor Systems

Hybrid FOSMC is a control strategy applied by combining different SMC subtypes with fractional operators according to the system’s dynamics and control requirements. This method combines the advantages of different SMC subtypes with fractional-order derivatives, thereby both improving control performance and providing a more flexible and stable response by considering the system’s past dynamics. When examining Table 3, it is seen that Hybrid FOSMC is used in studies [62,63,66,69,70,73,77,80,81], and among these studies, Hybrid FOSMC is applied to both internal and external loop structures in studies [66,73,80].

5.7. Hybrid SMC Approaches for Helicopter Systems

Hybrid SMC is an advanced control strategy formed by combining multiple subtypes under a single controller structure or by applying different controllers to each loop. In most of the studies [90,91,93,96,97,98,100,101] examined in Table 4, Hybrid SMC is used in the control of helicopter systems with two degrees of freedom, and it has been observed that the control parameters in these studies are generally determined by Metaheuristic (Genetic Algorithm, Particle Swarm Optimization, etc.) methods. In studies [88,103], respectively, Adaptive Super Twisting SMC and Observer-based Fixed-Time SMC methods are applied to the control of helicopter systems with three degrees of freedom. In study [99], the Integral Backstepping SMC combined with Neural Network method is used for the control of a six-degree-of-freedom helicopter system, and the control parameters are determined using the Trial and Error method.

5.8. Adaptive SMC Approaches for Helicopter Systems

Examining Table 4, it is observed that methods using the Adaptive structure as a Hybrid SMC component instead of classical Adaptive SMC are employed in the control of helicopter systems [88,93,96,98,100]. The use of Adaptive structures in such approaches, it enables the online optimization of gains by preventing them from being increased excessively against disturbing effects and model uncertainties. Thus, the control gain is only adapted to the required level, which reduces energy consumption by preventing unnecessary growth of the control signal. Additionally, the fact that the Adaptive structure does not require prior knowledge about uncertainties or disturbance limits reduces the system identification load in the design process and increases the ease of application of the method.

5.9. Super Twisting SMC Approaches for Helicopter Systems

When Table 4 is examined, methods using the Super Twisting controller as a Hybrid SMC component in the control of helicopter systems have been observed [88,90,93,95,100]. Approaches of this type are preferred because the Super Twisting algorithm smooths the discontinuous structure, making the derivative of the control signal continuous, and can directly address second-order sliding surface dynamics. Such a hybrid structure, significantly reducing the chattering effect enables high tracking performance to be maintained even under system uncertainties and disturbance effects.

5.10. Hybrid FOSMC Approaches for Helicopter Systems

When Table 5 is analyzed, it is observed that the application of FOSMC in helicopter system control is limited; the main reasons for this are the complex dynamic structure of the system, cross-interactions between axes, and computational difficulties in real-time applications. However, studies [106,109] show that hybrid FOSMC methods have been successfully applied and have yielded better results than the compared methods. Nevertheless, the failure to explicitly specify the fractional derivative operator used in the relevant studies constitutes a significant limitation. Since different fractional derivative definitions directly affect the system response and control performance, failure to specify the operator difficulties the repeatability of the method and comparative analyzes. This situation also difficulties the optimization and standardization of FOSMC designs for future research.

5.11. Fuzzy FOSMC Approaches for Helicopter Systems

In the Fuzzy FOSMC method, fuzzy logic smoothed the control signal by managing system uncertainties and nonlinear dynamics, while the fractional operator stabilizes this signal, thereby ensuring increased system performance and stability. Table 5 shows that there are a limited number of Fuzzy FOSMC applications in helicopter system control; the main reasons for this are the complexity of system dynamics, cross-interactions between axes, and the uncertainty of the fractional derivative operators used. However, some studies [106,110] have successfully applied this approach and achieved significant improvements in both control stability and robustness against parameter changes. In particular, study [110] used the Grünwald–Letnikov definition as the fractional operator and achieved successful results compared to the methods it compared.

5.12. Implementation Challenges and Computational Complexity

The findings of this review indicate that the majority of studies in the literature remain largely limited to simulation-based investigations. This trend reflects the technical constraints involved in transferring theoretical advantages to physical hardware. Although FOSMC approaches offer superior robustness and chattering reduction, their implementation on real-world UAV platforms introduces significant computational challenges. Unlike integer-order controllers, fractional-order operators possess an infinite memory property, which requires the entire history of the signal to be processed when computing fractional derivatives or integrals.

In practice, handling this cumulative data set poses a significant constraint on embedded flight controllers with limited processing power and memory. Consequently, the transition from simulation to physical systems necessitates the use of numerical approximation methods. However, employing such methods imposes additional computational burdens, potentially leading to latency in high-frequency control loops. As a result, ensuring the success of FOSMC strategies in real-time applications depends on establishing a trade-off between approximation accuracy and available computational resources. This balance is a decisive factor for transitioning these control methods from simulation to real-world applications. The relationship between the technical specifications, advantages, and implementation constraints of all reviewed strategies is presented as a comparative analysis in

Table 6. Comparative Analysis of Control Approaches.

Control Approach	Key Advantages	Key Disadvantages	Computational Cost	Tuning Complexity	Simulation Prevalence	Experimental Implementation
5.1. Hybrid SMC for Quadrotor	Maximum robustness and fault tolerance	Over-complicated design	High	High	High	Moderate
5.2. Adaptive SMC for Quadrotor	Full adaptation to payload variations	Risk of parameter drift	Moderate	Moderate	High	High
5.3. Terminal SMC for Quadrotor	Rapid finite-time convergence	Singularity risks near zero	Low	Moderate	High	Moderate
5.4. Terminal FOSMC for Quadrotor	High precision and finite-time convergence	Design complexity and parameter sensitivity	High	High	High	Low
5.5. Classical FOSMC for Quadrotor	Superior noise suppression capability	Memory effect and buffer overhead	Moderate	Moderate	Moderate	Moderate
5.6. Hybrid FOSMC for Quadrotor	Maximum tracking precision	Real-time coding difficulties	High	High	Moderate	Low
5.7. Hybrid SMC for Helicopter	Decoupling of inter-axis interactions	Synchronization of sub-controllers	High	High	Moderate	Moderate
5.8. Adaptive SMC for Helicopter	Online compensation of uncertainties	Transient regime instability risk	Moderate	Moderate	Moderate	Moderate
5.9. Super Twisting for Helicopter	Chattering-free control with reduced actuator wear	Difficulty in defining gain limits	Moderate	Low	High	High
5.10. Hybrid FOSMC for Helicopter	Optimal adaptation to non-linear structures	Latency at the application layer	High	High	Moderate	Low
5.11. Fuzzy FOSMC for Helicopter	Intelligent chattering suppression and disturbance  adaptation	Maximum processor load occurs due to the dual-layer structure	High	Moderate	Moderate	Moderate

.

The comparative analysis presented in Table 6 clearly illustrates the critical trade-off between theoretical control performance and practical applicability. In particular, while hybrid FOSMC structures offer the highest tracking precision in simulation environments, their experimental applicability remains low due to high computational costs and the processor load resulting from the infinite memory effect of fractional-order operators.

Based on the technical parameters in Table 6, approaches such as SMC, Terminal SMC, and Integral SMC provide fundamental hardware compatibility and ease of implementation owing to their low computational overhead and manageable tuning complexity. However, Adaptive SMC and Super Twisting SMC stand out as the most optimized solutions for real-time UAV missions, as they integrate this low-cost architecture with critical operational advantages such as parameter estimation and chattering suppression, respectively. This synthesis provides a clear overview of the current technical state between complex theoretical approaches in the literature and the hardware constraints of physical systems.

6. Conclusions

6.1. Findings

The SMC and FOSMC approaches used in the control of Quadrotor and Helicopter systems within rotary-wing UAV systems have been comprehensively examined based on methods proposed in previous studies. This comprehensive evaluation covers the research results of approximately 77 studies published between 2015 and 2025. First, SMC and FOSMC-based control methods applied to Quadrotor systems were analyzed, followed by the application of similar control strategies to helicopter systems. The findings reveal a steady upward trend in both SMC and FOSMC studies over the years for Quadrotor systems. In particular, fractional-order methods have become increasingly preferred in recent years. In helicopter systems, however, the trend is more volatile. Although there have been occasional increases in SMC studies, the overall research profile is unstable. It is understood that FOSMC studies for helicopters remain quite limited and have the potential to form a broad field of research. In general, the rotary-wing UAV literature shows that Quadrotor models have become the dominant research platform; however, in helicopter systems, fractional-order control methods in particular are still understudied, presenting a significant research gap. Therefore, this review is expected to draw the attention of researchers, facilitate their understanding of the most recent SMC and FOSMC control strategies used in rotary-wing UAV systems, and guide future work in this area.

6.2. Future Perspectives

• Quadrotor Systems

• The fact that a very high proportion ($83.02\%$) of the reviewed quadrotor studies remain solely in simulation environments is considered to stem from the heavy computational load created by the infinite memory effect of FOSMC algorithms on real-time embedded hardware. This makes the development of efficient approximation algorithms that minimize processing load while maintaining hardware precision the most prioritized technical need in the literature.
• Given that Trial-and-Error methods are used at a rate of $36.67\%$ in determining controller parameters, and considering the limitations of current metaheuristic optimization techniques; the integration of Deep Reinforcement Learning-based structures capable of instantly updating parameters and sliding manifold coefficients according to changing flight conditions into FOSMC is a significant research focus for increasing adaptation capability in challenging scenarios.
• Future research should move beyond the rigid-body assumption and focus on high-fidelity modeling approaches that fill theoretical gaps by incorporating flexible body dynamics and nonlinear aerodynamic ground effects into FOSMC structures.
• According to the literature review conducted within the scope of this study, the Caputo–Fabrizio (CF) definition has not been encountered in fractional-order controller applications. This offers an original research opportunity to investigate the effects of the non-singular kernel property of the CF definition on chattering suppression and control accuracy in UAV systems.
• While the integration of Super Twisting FOSMC offers a significant opportunity for chattering suppression, the development of mathematical methods capable of handling non-smooth disturbances with unlimited rates of change and automatically adjusting controller parameters remains a significant theoretical deficiency.

• Helicopter Systems

• The fact that $79.17\%$ of helicopter studies are limited to 2-DOF models constitutes a fundamental research focus for advanced MIMO-FOSMC architectures capable of managing complex cross-couplings and disturbance torque interactions between the main and tail rotors.
• Considering the high parameter sensitivity and nonlinear structure of helicopters; supporting FOSMC with hybrid intelligent algorithms, such as metaheuristic optimization or artificial neural networks, in online parameter tuning processes is a critical area of study for ensuring performance stability.
• The failure to specify the fractional operator type in $40\%$ of the reviewed articles creates uncertainty regarding optimal performance in different flight regimes. Establishing theoretical selection criteria for these systems will provide a significant contribution to the literature.
• In helicopter systems prone to failures due to mechanical complexity, the development of FOSMC-based Fault-Tolerant Control strategies and the theoretical guarantee of system stability in critical scenarios is a prioritized need.