Abstract:A persistent challenge in tasks involving large-scale dynamical systems, such as state estimation and error reduction, revolves around processing the collected measurements. Frequently, these data suffer from the curse of dimensionality, leading to increased computational demands in data processing methodologies. Recent scholarly investigations have underscored the utility of delineating collective states and dynamics via moment-based representations. These representations serve as a form of sufficient statistics for encapsulating collective characteristics, while simultaneously permitting the retrieval of individual data points. In this paper, we reshape the Kalman filter methodology, aiming its application in the moment domain of an ensemble system and developing the basis for moment ensemble noise filtering. The moment system is defined with respect to the normalized Legendre polynomials, and it is shown that its orthogonal basis structure introduces unique benefits for the application of Kalman filter for both i.i.d. and universal Gaussian disturbances. The proposed method thrives from the reduction in problem dimension, which is unbounded within the state-space representation, and can achieve significantly smaller values when converted to the truncated moment-space. Furthermore, the robustness of moment data toward outliers and localized inaccuracies is an additional positive aspect of this approach. The methodology is applied for an ensemble of harmonic oscillators and units following aircraft dynamics, with results showcasing a reduction in both cumulative absolute error and covariance with reduced calculation cost due to the realization of operations within the moment framework conceived.

Ensemble Kalman filtering for nonlinear systems with multiple delayed measurements

Applying a Novel Extended Kalman Filter to Missile–target Interception with APN Guidance Law: A Benchmark Case Study

Ensemble Consider Kalman Filtering

Extended target tracking for high resolution sensor based ensemble Kalman filters

Kalman Filtering for Continuous-Time Systems with Multiple Delayed Measurements

Set-Membership Based Hybrid Kalman Filter for Nonlinear State Estimation under Systematic Uncertainty.

A Novel Extended Kalman Filter for A Class of Nonlinear Systems

A moment-based Kalman filtering approach for estimation in ensemble systems

Extended Kalman Filtering with Stochastic Nonlinearities and Multiple Missing Measurements

Extended Target Tracking Using An Imm Based Nonlinear Kalman Filters

Nonlinear Measurement Function in the Ensemble Kalman Filter

Cubature Ensemble Kalman Filter for Highly Dimensional Strongly Nonlinear Systems.

Unscented Kalman-Bucy filtering for nonlinear continuous-time systems with multiple delayed measurements

Universal delayed kalman filter with measurement weighted summation for the linear time invariant system

Adaptive Error Covariances Estimation Methods for Ensemble Kalman Filters

Hybrid Kalman Filtering Algorithm with Stochastic Nonlinearities and Multiple Missing Measurements

Nonlinear Kalman Filtering with Numerical Integration

A New Nonlinear State Estimator Using the Fusion of Multiple Extended Kalman Filters

Kalman filtering for multiple time-delay systems

Adaptive Multinoulli-Based Kalman Filter with Randomly Unknown Delayed and Lost Measurements

Distributed Finite-Horizon Extended Kalman Filtering for Uncertain Nonlinear Systems