In the field of remote sensing, tracking highly maneuvering targets is challenging due to its rapidly changing patterns and uncertainties, particularly under non-Gaussian noise conditions. In this paper, we consider the problem of tracking highly maneuvering targets without using preset parameters in non-Gaussian noise. We propose a multi-domain intelligent state estimation network (MIENet). It consists of two main models to estimate the key parameter for the Unscented Kalman Filter, enabling robust tracking of highly maneuvering targets under various intensities and distributions of observation noise. The first model, called fusion denoising model (FDM), is designed to eliminate observation noise by enhancing multi-domain feature fusion. The second model, called parameter estimation model (PEM), is designed to estimate key parameters of target motion by learning both global and local motion information. Additionally, we design a physically constrained loss function (PCLoss) that incorporates physics-informed constraints and prior knowledge. We evaluate our method on radar trajectory simulation and real remote sensing video datasets. Simulation results on the LAST dataset demonstrate that the proposed FDM can reduce the root mean square error (RMSE) of observation noise by more than 60\%. Moreover, the proposed MIENet consistently outperforms the state-of-the-art state estimation algorithms across various highly maneuvering scenes, achieving this performance without requiring adjustment of noise parameters under non-Gaussian noise. Furthermore, experiments conducted on the real-world SV248S dataset confirm that MIENet effectively generalizes to satellite video object tracking tasks.

In the field of remote sensing, tracking highly maneuvering targets is challenging due to its rapidly changing patterns and uncertainties, particularly under non-Gaussian noise conditions. In this paper, we consider the problem of tracking highly maneuvering targets without using preset parameters in non-Gaussian noise. We propose a multi-domain intelligent state estimation network (MIENet). It consists of two main models to estimate the key parameter for the Unscented Kalman Filter, enabling robust tracking of highly maneuvering targets under various intensities and distributions of observation noise. The first model, called a fusion denoising model (FDM), is designed to eliminate observation noise by enhancing multi-domain feature fusion. The second model, called a parameter estimation model (PEM), is designed to estimate key parameters of target motion by learning both global and local motion information. Additionally, we design a physically constrained loss function (PCLoss) that incorporates physics-informed constraints and prior knowledge. We evaluate our method on radar trajectory simulation and real remote sensing video datasets. Simulation results on the LAST dataset demonstrate that the proposed FDM can reduce the root mean square error (RMSE) of observation noise by more than 60%. Moreover, the proposed MIENet consistently outperforms the state-of-the-art state estimation algorithms across various highly maneuvering scenes, achieving this performance without requiring adjustment of noise parameters under non-Gaussian noise. Furthermore, experiments conducted on the real-world SV248S dataset confirm that MIENet effectively generalizes to satellite video object tracking tasks.

Pattern recognition of maneuvering target is challenging due to its rapidly changing patterns and uncertainties. Short-term trajectories are essential for capturing dynamic motion changes. However, existing deep-learning based methods rely on long-term trajectories to extract more comprehensive information, making them less effective with limited data. In this paper, we focus on recognizing short-term trajectories and propose a multilevel integrated framework that narrows the latent search space. Our approach employs a two-level classification with meta-learning (MetaMLIR) to recognize target motion patterns by learning multivariate features from noisy observations. Specifically, the first-level classification generates preliminary scores derived from multiple heterogeneous classifiers with a balanced integration strategy. The second-level classification uses logistic regression with meta-learning to share knowledge from the first-level outputs and provide the final classification result for each motion behavior. Additionally, we develop a target motion trajectory dataset (namely, TMT) to conduct comprehensive performance evaluation. Experimental results on our self-developed TMT dataset and the real-world NCLT dataset demonstrate that the proposed method achieves enhanced motion-pattern recognition performance across both non-pattern-switching and pattern-switching scenarios.