Piezo-driven fast steering mirrors (PFSMs) are widely employed in high-precision beam steering and accurate tracking applications. However, the inherent hysteresis nonlinearity of piezoelectric actuators significantly degrades tracking accuracy. To address the challenges posed by dynamic hysteresis nonlinearity, this study proposes an improved non-singular fast terminal sliding mode control strategy. The proposed method integrates a non-singular fast terminal sliding surface and introduces an adaptive function in the reaching law to enhance response speed and improve control robustness. Additionally, the strategy incorporates an extended state observer (ESO) and an inverse model-based feedforward compensation mechanism. Specifically, the feedforward compensation based on the inverse model aims to offset hysteresis effects, while the ESO provides a real-time estimation of the total system disturbance to mitigate the impact of external disturbances and unmodeled hysteresis. Experimental results demonstrate that the proposed method effectively compensates for the hysteresis nonlinearity of PFSMs, improves disturbance rejection performance, and enhances position control accuracy.

Under ground-based observation geometry, infrared radiant-intensity sequences of space objects are jointly influenced by object micromotion, thermal radiation, time-varying viewing conditions, and atmospheric propagation. Existing simulation studies often prescribe the line of sight or simplify the coupling between viewing geometry and atmospheric attenuation, which limits long-duration ground-based sequence analysis. This study develops a physics-informed framework for generating atmosphere-attenuated infrared radiant-intensity sequences of space objects undergoing precession or tumbling. The framework reconstructs observation geometry from azimuth–elevation–range trajectories, updates facet normals through a unified micromotion attitude model, computes visible projected area and transient facet temperature, and incorporates MODTRAN-derived elevation-dependent atmospheric transmittance. Using this framework, we construct IRPeriodic, an eight-class simulated dataset for long-duration univariate time-series classification. We further propose LPD-Net, which integrates large-kernel residual feature extraction, prototype-guided dynamic temporal alignment, and differential periodic representation to capture long-range waveform morphology, sample-dependent temporal correspondence, and segment-level local variation. On IRPeriodic, LPD-Net achieves an accuracy of 0.8618 ± 0.0057, a macro-F1 of 0.8615 ± 0.0061, and a Matthews correlation coefficient of 0.8426 ± 0.0065, outperforming the evaluated neural-network and ROCKET-type baselines. Ablation and synthetic-noise sensitivity analyses indicate that the performance gain is mainly associated with long-context feature extraction, with additional improvements from dynamic alignment and differential periodic statistics. Auxiliary experiments on selected public UCR datasets suggest that the representation is also competitive for univariate time-series classification. These results demonstrate the effectiveness of LPD-Net on the proposed physics-informed benchmark for long-duration ground-based infrared radiant-intensity sequence classification.