Extended Kalman Filtering on Stiefel Manifolds

Jordi-Lluís Figueras and Aron Persson and Lauri Viitasaari · 2025

Read Extended Kalman Filtering on Stiefel Manifolds by Jordi-Lluís Figueras and Aron Persson and Lauri Viitasaari (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-02682.txt.

Introduction

Tracking a moving object whose readings live on a curved space calls for special filters. Jordi-Lluis Figueras, Aron Persson, and Lauri Viitasaari extend the well-known extended Kalman filter to the Stiefel manifold in a 2025 paper. That manifold covers readings like points on a sphere and sets of perpendicular axes, both common in tracking work.

The authors test their filter on two settings: points moving on a sphere, and four-by-two matrices with perpendicular columns. In both cases the filter's estimates beat the raw readings by a wide margin. The work gives engineers a general tool for cleaning up noisy data that naturally takes values on these curved spaces, rather than flat ones.

Catalog record

CREATORJordi-Lluís Figueras and Aron Persson and Lauri Viitasaari
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-02682

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Extended Kalman Filtering on Stiefel Manifolds Jordi-Lluís Figueras and Aron Persson and Lauri Viitasaari Abstract . A generalisation of the extended Kalman filter for Stiefel manifold-valued measurements is presented. We provide simulations on the 2-sphere and the space of orthogonal 4-by-2 matrices which show significant improvement of the Extended Kalman Filter compared to only relying on raw measurements. 1. Introduction 1.1. Examples of projected dynamics There are multiple situations where projected dynamics is inherently tied to gaining information. We shall begin by giving simple examples of two such situations in order to broaden the perspective before we go into the specifics relating to applications of Stiefel manifold-valued statistics in radiology. Consider a radioactive dust particle travelling through an open space. Suppose also there is a moderate wind going through the area and we can only measure the direction of the incoming radiation. In physics one often models the movement of dust particles as if they travel like Brownian motion with a drift (the drift here comes from the wind). Therefore the dust particle does not travel deterministically and there is an uncertainty inherent to the position of the dust particle. Moreover, the measured radiation incoming into our sensor might have a somewhat perturbed inclination error as it collides with air particles in between the dust particle and the sensor. Hence, there is an uncertainty of the incoming direction of the radiated particles as well. In this situation the dust particle is travelling across \mathbb{R}^{3} and our measurements are taken on the sphere \mathbb{S}^{2} . The information obtained is (non-linearly) projected onto the sphere. Another plausible scenario would be a circular particle accelerator. As a charged particle is centripetally accelerated through electro-magnetic forces we measure its position in the accelerator using sensors throughout the walls of the accelerator. As the particle travelling through the accelerator is a quantum particle, the position of the particle is inherently probabilistic. At the same time the sensors are not perfect and there is uncertainty in the measurements. Here the particle moves inside a tubular neighbourhood of a circle and the measurements happen on the torus \mathbb{T}^{2} . In MIMO radio-systems for cellular networks a configuration of n antennas interfere constructively and a directed radio signal is produced with high energy efficiency. This signal is then received at and responded by k receivers. These receivers move around, in part, in a predictive way (a person moving in a certain direction will probably move in that direction for a while) and, in part, with some uncertainty. Hence, these receivers may be modelled as stochastic processes with some dynamical component. Moreover, the antenna configurations can be modelled as being a vector in \mathbb{R}^{n} for each receiver. In total we consider the state as an n by k matrix. However, as these configurations do not depend on the total strength of the currents, only their relative strength and phase, so measurements are realised as n by k orthogonal matrices, see [ 6 , 15 , 16 ] . These realisations may be computed by only retaining the orthogonal part, Q , in the polar decomposition of the n by k matrix. This Q can be computed so that it is a projection onto the the space of n by k orthogonal matrices, i.e. the Stiefel manifold \St_{n,k} . These measurements are assumed to be noisy and the measured optimal configuration of the antennae has some uncertainty. All these three situations illuminates a mathematical problem: an object moves around with a drift in some (possibly an open subset of) vector space, see Figure 2 for a picture of these situations. If one performs measurements on a non-linear space, i.e. if x,y\in X , then it is generally not true that x+y\in X . How would one filter out this information using both the known drift of the system and the measurements? In all the above examples the non-linear spaces are manifolds which have sufficient geometrical structure which allows for measurement of distances between points. Using distances and curves on these manifolds it is then possible to weight the prediction together with the measurement. In a nutshell, working on the original filtering equations developed in [ 7 , 8 , 9 ] we will filter the predicted point and the measurement by using a curve that connects these two points. Then one can obtain the filtered mean by taking a weighted average along the curve. This paper can be viewed as part of the vast field of directional statistics. The area of directional statistics has found applications in medicine, see [ 2 , 3 , 10 ] , in meteorology, see [ 12 , 13 , 14 ] , and in robotics, see [ 11 , 18 ] , to name a few. 1.2. Organisation of the paper Section 2 explains the foundational concepts for the problem at hand. In Section 3 we give an exposition on how to use the extended Kalman filter when the measurements are points on the Stiefel manifold \St_{n,k} . The reader with little interest of the geometrical meaning behind the Extended Kalman filter on \St_{n,k} may jump directly to Algorithm 2 at the end of this section. In Sections 4 and 5 we give results from simulations for the 2-sphere and \St_{4,2} , respectively. In Section 6 we discuss some limitations and we present some potential drawbacks of the extended Kalman filter on Stiefel manifolds. 2. Setting Suppose an object X_{t} is represented as an n by k matrix, that is X_{t}\in\mathbb{R}^{n\times k} . As t increases we suppose it moves around with known dynamics and with non-zero noise (therefore the movement is not fully deterministic). This may be mathematically represented as an SDE (Stochastic Differential Equation) (2.1) \differential X_{t}=AX_{t}\differential t+\nu\differential B_{t},\qquad X_{0}\eqdist N(\mu_{0},\nu_{0}^{2}\id_{\mathbb{R}^{n\times k}}). Here \id_{\mathbb{R}^{n\times k}} denotes the identity map from the space of n by k matrices to itself, i.e. for given Y\in\mathbb{R}^{n\times k} , \id_{\mathbb{R}^{n\times k}}(Y)=Y . More explicitly, if Y_{\text{vec}}\in\mathbb{R}^{nk} is the vectorisation of an n times k matrix Y , then if \id_{\mathbb{R}^{n\times k}}^{\text{vec}} is the corresponding vectorized linear map, it is the nk\times nk identity matrix. The initial value X_{0} is the starting (normal) distribution of X_{t} and A\in\mathbb{R}^{n\times n} describes the dynamics which is the infinitesimal displacement over time. The term \differential B_{t} is a formal stochastic differential of the Brownian motion which we scale by a factor \nu . After a time step \delta t>0 and if A=0 , then solving ( 2.1 ) one would obtain X_{t}=X_{0}+\nu\int_{0}^{\delta t}\differential B_{t} which has variance matrix (\nu_{0}^{2}+\nu^{2}\delta t)\id_{\mathbb{R}^{n\times k}} and constant mean \mu_{0} . More generally, if A is anti-symmetric, i.e. A^{T}=-A , then one obtains the full solution to ( 2.1 ) as X_{t}\eqdist N(\exp_{\mathrm{M}}(tA)\mu_{0},(\nu_{0}^{2}+t\nu^{2})\id_{\mathbb{R}^{n\times k}}). Here \exp_{\mathrm{M}}:\mathbb{R}^{n\times n}\rightarrow\mathbb{R}^{n\times n} denotes the matrix exponential defined as \exp_{\mathrm{M}}(A)=\sum_{j=0}^{\infty}\frac{A^{j}}{j!}, where A^{j} denotes j -times matrix multiplication and A^{0}=I_{n\times n} , the n\times n identity matrix. (Recall that this series is absolutely convergent for any A ). Remark 2.1 . Note that this solution is only true when the factor \nu is a scalar (for which we then call the noise isotropic). If the factor \nu is not a scalar then there will appear more time dependent terms in the

Read the complete text