TL;DR
This paper introduces a novel Generalized High-Order Extended State Observer (GHOESO) to precisely estimate and suppress periodic disturbances in airborne electro-optical stabilized platforms. By incorporating known disturbance frequencies, the GHOESO significantly outperforms traditional observers, leading to improved stabilization accuracy in challenging flight conditions.
What Problem Does This Paper Address?
Airborne electro-optical stabilized platforms are crucial for reconnaissance, targeting, and tracking from aircraft. These platforms must maintain a stable line-of-sight (LOS) despite the aircraft's motion and environmental disturbances like wind, friction, and vibrations. Traditional control methods struggle with these disturbances because they often require precise mathematical models of the system and disturbances, which are difficult to obtain in dynamic airborne environments. Active Disturbance Rejection Control (ADRC) offers a model-free approach, with the Extended State Observer (ESO) being its core component for estimating and compensating disturbances. However, standard ESOs are only effective for constant or slowly varying disturbances and perform poorly when faced with periodic, time-varying disturbances, which are common in airborne applications. This limitation hinders the platform's stabilization accuracy and overall performance.
Methods
The paper proposes a Generalized High-Order Extended State Observer (GHOESO) to address the limitations of traditional ESOs in handling periodic disturbances. The core idea is to leverage prior knowledge of the disturbance's frequency.
- Traditional ESO: The paper first reviews the standard ESO, which treats internal uncertainties and external disturbances as a single "lumped disturbance" and estimates it as an extended state. This approach is effective for constant disturbances but struggles with dynamic ones.
- GHOESO Design: The GHOESO extends the ESO by incorporating the known frequency of sinusoidal disturbances. It models the disturbance's second-order derivative based on its known frequency (e.g., $d(t) = -w^2 d(t)$ for a sinusoidal disturbance). This allows the observer to track and estimate the disturbance more accurately, especially when it's periodic. The GHOESO's state vector is expanded to include not only the system states but also the disturbance and its derivatives up to a certain order, informed by the disturbance's known frequency.
- Stability Analysis: The paper provides a theoretical analysis of the GHOESO's stability, demonstrating that the observation error converges to a bounded interval, ensuring reliable disturbance estimation.
Datasets
Not discussed in the paper. The paper relies on simulated data derived from a mathematical model of an airborne photoelectric stabilization platform and experimental data collected from a laboratory setup.
Key Insights and Ideas
- Frequency is Key: The most significant insight is that incorporating known disturbance frequencies into the observer design dramatically improves its ability to track and reject periodic disturbances. This moves beyond simply increasing observer gain, which can lead to other issues.
- Beyond Constant Disturbances: The paper clearly demonstrates that traditional ESOs are fundamentally limited by their assumption of slowly varying or constant disturbances. Periodic signals, even at moderate frequencies, pose a significant challenge.
- GHOESO's Robustness: The GHOESO shows remarkable resilience to increasing disturbance frequencies, maintaining high accuracy where traditional ESOs falter. This is crucial for airborne platforms facing a wide range of dynamic environmental conditions.
- Experimental Validation: The successful experimental validation on a hexapod swing table provides strong evidence for the GHOESO's practical applicability and superiority over traditional methods in a realistic scenario.
Key Results
- 1 Hz Sinusoidal Disturbance: The GHOESO achieved a disturbance error of less than 0.01 and a velocity fluctuation of 0.002°/s, compared to the ESO's error of 0.3 and fluctuation of 0.11°/s.
- 3 Hz Sinusoidal Disturbance: The GHOESO maintained a disturbance error of less than 0.01 and a velocity fluctuation of 0.002°/s. In contrast, the ESO's error increased to 2.2 and velocity fluctuation to 0.2°/s, highlighting its degradation at higher frequencies.
- Compound Disturbance: Even under complex superimposed disturbances, the GHOESO outperformed the ESO. The GHOESO achieved a disturbance error of 1.6 and a velocity fluctuation of 0.045°/s, while the ESO had an error of 4 and fluctuation of 0.15°/s.
- Experimental Results: The GHOESO reduced the total miss distance by 53.93% at 1°/1 Hz compared to the ESO, demonstrating significant improvements in stabilization accuracy.
Limitations and Future Work
- Known Frequency Assumption: The primary limitation is the requirement for known disturbance frequencies. The GHOESO's effectiveness is directly tied to this prior information. If disturbance frequencies are unknown or rapidly changing, this specific GHOESO design may not be optimal.
- Complex Disturbances: While the GHOESO performs better than the ESO under compound disturbances, its performance is still compromised, particularly at the peak of the disturbance wave.
- Future Work: The authors suggest further research into handling unknown or time-varying disturbance frequencies, potentially by combining the GHOESO with adaptive frequency estimation techniques. Investigating the GHOESO's performance with more complex, non-sinusoidal disturbances and exploring its application to other stabilization systems would also be valuable.
Summary updated