RBFNN-Based Singularity-Free Terminal Sliding Mode Control for Uncertain Quadrotor UAVs

Area of Science:

• Robotics and Control Systems
• Aerospace Engineering
• Artificial Intelligence

Background:

• Quadrotor unmanned aerial vehicles (QUAVs) are susceptible to performance degradation due to inertia uncertainties and external disturbances.
• Traditional control methods often struggle with singularities and require precise system models, limiting their applicability.
• Ensuring robust attitude and angular velocity control is critical for safe and effective QUAV operation.

Purpose of the Study:

• To propose a singularity-free terminal sliding mode (SFTSM) control scheme for QUAVs.
• To address challenges posed by inertia uncertainties and external disturbances without requiring prior system knowledge.
• To achieve finite-time convergence and robust stability for QUAV attitude control.

Main Methods:

• Construction of a singularity-free terminal sliding mode surface (SFTSMS) for finite-time convergence.
• Design of an adaptive finite-time control with an auxiliary function to circumvent singularity issues.
• Integration of a radial basis function neural network (RBFNN) and an extended state observer (ESO) for disturbance estimation.

Main Results:

• The proposed SFTSM control scheme ensures finite-time uniform ultimate boundedness (FTUUB) for attitude and angular velocity errors.
• Numerical simulations demonstrate the effectiveness and satisfactory performance of the control strategy.
• The method successfully estimates unknown disturbances, eliminating the need for detailed system model information.

Conclusions:

• The developed SFTSM control scheme offers a robust and effective solution for controlling QUAVs in the presence of uncertainties and disturbances.
• The integration of RBFNN and ESO enhances the adaptability and accuracy of the control system.
• This approach contributes to improved stability and performance of autonomous aerial vehicles.