Abstract:A stable square-root approach has been recently proposed for the unscented Kalman filter (UKF) and fifth-degree cubature Kalman filter (5D-CKF) as well as for the mixed-type methods consisting of the extended Kalman filter (EKF) time update and the UKF/5D-CKF measurement update steps. The mixed-type estimators provide a good balance in trading between estimation accuracy and computational demand because of the EKF moment differential equations involved. The key benefit is a consolidation of reliable state mean and error covariance propagation by using delicate discretization error control while solving the EKF moment differential equations and an accurate measurement update according to the advanced UKF and/or 5D-CKF filtering strategies. Meanwhile the drawback of the previously proposed estimators is an utilization of sophisticated numerical integration scheme with the built-in discretization error control that is, in fact, a complicated and computationally costly tool. In contrast, we design here the mixed-type methods that keep the same estimation quality but reduce a computational time significantly. The novel estimators elegantly utilize any MATLAB-based numerical integration scheme developed for solving ordinary differential equations (ODEs) with the required accuracy tolerance pre-defined by users. In summary, a simplicity of the suggested estimators, their numerical robustness with respect to roundoff due to the square-root form utilized as well as their estimation accuracy due to the MATLAB ODEs solvers with discretization error control involved are the attractive features of the novel estimators. The numerical experiments are provided for illustrating a performance of the suggested methods in comparison with the existing ones.

Numerically robust square root implementations of statistical linear regression filters and smoothers

New square-root smoothing algorithms

Numerically Robust Fixed-Point Smoothing Without State Augmentation

A Square-Root Kalman Filter Using Only QR Decompositions

Square-root RTS smoothing algorithms

Square-root filtering via covariance SVD factors in the accurate continuous-discrete extended-cubature Kalman filter

MATLAB-based general approach for square-root extended-unscented and fifth-degree cubature Kalman filtering methods

Robust Recursive Filtering and Smoothing

Improved Discrete-Time Kalman Filtering within Singular Value Decomposition

Robust and Trend Following Student's t Kalman Smoothers

On the stable Cholesky factorization-based method for the maximum correntropy criterion Kalman filtering

On Regularized Square-root Regression Problems: Distributionally Robust Interpretation and Fast Computations

Fast and accurate approximation algorithms for computing floating point square root

Robust Normalized Least Mean Absolute Third Algorithms.

Posterior linearisation smoothing with robust iterations

On The Robustness Of Numerical Algorithms For Linear Systems And Signal Processing In Finite Precision Arithmetic

Continuous-discrete unscented Kalman filtering framework by MATLAB ODE solvers and square-root methods

A Study of the Robustness of Iterative Methods for Linear Systems

Factored-form Kalman-like implementations under maximum correntropy criterion

Rational function approximation with normalized positive denominators

Square Root Unscented Particle Filter With Application To Angle-Only Tracking