Inertial stabilization platforms (ISPs) on unmanned aerial vehicles (UAVs) are critical for clear imaging and accurate measurement of ground/water targets. However, ISPs often suffer from performance degradation due to complex disturbances, especially the dominant periodic disturbances. Traditional extended state observers (ESOs) struggle to effectively handle these time-varying periodic disturbances, limiting line-of-sight stabilization accuracy. To overcome this limitation, this paper proposes a novel third-order Adaptive Generalized Extended State Observer (3-AGESO). This method integrates disturbance frequency information directly into the observer's state expansion. A tracking differentiator, dynamically tuned via adaptive control parameters, reconstructs the disturbance frequency in real-time, ensuring robustness against frequency variations. The experiments show that the velocity fluctuations at the disturbance frequency of 1 Hz are reduced from 0.075 to 0.005°/s. The visual axis jitter of the ISP is reduced from 15.11 pixels to 3.77 pixels. These results demonstrate that the 3-AGESO effectively addresses the limitations of traditional ESOs in handling periodic disturbances. Consequently, it significantly enhances ISP stabilization and improves the precision of UAV-based imaging and measurement.
Wang et al. (2025) studied this question.