Novel anti-saturation adaptive sliding-mode-assisted disturbance observer for aerial imaging systems.

Abstract

Aerial imaging systems are subject to complex multisource disturbances, including high-frequency engine vibrations and abrupt aerodynamic turbulence, which critically hinder line-of-sight stabilization. Moreover, input saturation nonlinearities imposed by stringent onboard hardware constraints severely limit the capability of the system to reject highly dynamic perturbations. To address these challenges, this study proposes a novel anti-saturation adaptive sliding-mode-assisted disturbance observer. The proposed scheme integrates a dual-strategy anti-saturation mechanism. First, a dynamic unsaturation-scale factor mathematically quantifies the residual driving capacity of the actuator to actively regulate the adaptive sliding-mode gain in real time. Second, a describing-function-based constraint is strictly embedded into the exact observer error dynamics. This dual approach intelligently manages the control effort within absolute physical limits, preventing integral windup, gain over-regulation, and catastrophic observer divergence during deep saturation. Through rigorous Lyapunov stability analysis, all closed-loop signals are proven to exhibit global uniform ultimate boundedness. Finally, comparative experiments conducted on an equivalent gimbal platform validated the superiority of the proposed controller in achieving both high dynamic disturbance rejection and high-precision imaging in adverse aerial environments. The proposed algorithm outperformed existing anti-disturbance methods in terms of convergence speed and dynamic performance, thereby validating its robustness.