π€ AI Summary
This work addresses the limitations of the conventional unscented Kalman filter (UKF) under time-varying dynamics and heavy-tailed non-Gaussian noise, which stem from its reliance on static parameterization. To overcome this, we propose a memory-augmented meta-learning framework that, for the first time, formulates sigma-point weight generation as a hyperparameter optimization problem. A recurrent context encoder compresses historical innovation sequences, and a policy network dynamically synthesizes weights for the mean and covariance to adaptively balance trust between prediction and measurement. This approach departs from fixed or heuristic tuning paradigms, enabling end-to-end training and out-of-distribution generalization. Evaluated on maneuvering target tracking tasks, the method significantly outperforms standard UKF baselines, demonstrates enhanced robustness to flicker noise, and generalizes effectively to unseen dynamic scenarios.
π Abstract
The Unscented Kalman Filter (UKF) is a ubiquitous tool for nonlinear state estimation; however, its performance is limited by the static parameterization of the Unscented Transform (UT). Conventional weighting schemes, governed by fixed scaling parameters, assume implicit Gaussianity and fail to adapt to time-varying dynamics or heavy-tailed measurement noise. This work introduces the Meta-Adaptive UKF (MA-UKF), a framework that reformulates sigma-point weight synthesis as a hyperparameter optimization problem addressed via memory-augmented meta-learning. Unlike standard adaptive filters that rely on instantaneous heuristic corrections, our approach employs a Recurrent Context Encoder to compress the history of measurement innovations into a compact latent embedding. This embedding informs a policy network that dynamically synthesizes the mean and covariance weights of the sigma points at each time step, effectively governing the filter's trust in the prediction versus the measurement. By optimizing the system end-to-end through the filter's recursive logic, the MA-UKF learns to maximize tracking accuracy while maintaining estimation consistency. Numerical benchmarks on maneuvering targets demonstrate that the MA-UKF significantly outperforms standard baselines, exhibiting superior robustness to non-Gaussian glint noise and effective generalization to out-of-distribution (OOD) dynamic regimes unseen during training.