Recent Advance In Computing Kolmogorov Equation Arising From Nonlinear Filtering

Sst Yau,X Wu,Kp Lin

DOI: https://doi.org/10.1109/CDC.1998.757918

1998-01-01

Abstract:It is well known that Kolmogorov equation is a fundamental equation in Applied Science, especially in Electrical Engineering. In this paper, we present two closed form solutions to Kolmogorov equation which plays an essential role in nonlinear filtering.

What problem does this paper attempt to address?

Special Solutions to Some Kolmogorov Equations Arising from Cubic Sensor Problems.

Ruxu Du,Lixing Jia,Stephen S.-T. Yau

DOI: https://doi.org/10.4310/cis.2007.v7.n2.a6

2007-01-01

Abstract:1. Introduction. Ever since the technique of the Kalman-Bucy filter was pop- ularized, there has been an intense interest in developing nonlinear filtering theory. Basically we have a signal or state process x = {xt} which is usually not observable. What we can observe is a related process y = {yt}. The goal of nonlinear filtering is to determine the conditional expectation of the form E(�(xt) : ys,0 � st) whereis any C ∞ function or even better to compute the entire conditional probability density �(t,x) of xt given the observation history {ys : 0 � st}. In practical applications, it is preferable that the computation of conditional probability density be preformed recursively in terms of a statistic � = {�t}, which can be updated by using only the latest observations. In some cases, �t is computable with a finite system of differential equations driven by y. This leads to the ideal notion of finite dimensional recursive filter. By definition such a filter is a system: dt = �(�t)dt + p X i=1 �i(�t)dyit
Finite-dimensional Filters with Nonlinear Drift. V: Solution to Kolmogorov Equation Arising from Linear Filtering with Non-Gaussian Initial Condition

ZG Liang,SST Yau,ST Yau

DOI: https://doi.org/10.1109/7.625130

IF: 3.491

1997-01-01

IEEE Transactions on Aerospace and Electronic Systems

Abstract:Despite its usefulness, the Kalman-Bucy filter is not perfect. One of its weaknesses is that it needs a Gaussian assumption on the initial data. Recently Yau and Yau introduced a new direct method to solve the estimation problem for linear filtering with non-Gaussian initial data. They factored the problem into two parts: (1) the on-line solution of a finite system of ordinary differential equations (ODEs), and (2) the off-line calculation of the Kolmogorov equation. Here we derive an explicit closed-form solution of the Kolmogorov equation. We also give some properties and conduct a numerical study of the solution.
Wavelet-Galerkin Method for the Kolmogorov Equation

Zhigang Liang,Stephen S.-T. Yau

DOI: https://doi.org/10.1016/j.mcm.2003.07.016

2004-01-01

Abstract:It is well known that the Kolmogorov equation plays an important role in applied science. For example, the nonlinear filtering problem, which plays a key role in modern technologies, was solved by Yau and Yau [1] by reducing it to the off-line computation of the Kolmogorov equation. In this paper, we develop a theorical foundation of using the wavelet-Galerkin method to solve linear parabolic P.D.E. We apply our theory to the Kolmogorov equation. We give a rigorous proof that the solution of the Kolmogorov equation can be approximated very well in any finite domain by our wavelet-Galerkin method. An example is provided by using Daubechies D 4 scaling functions.
Explicit Solution of A Kolmogorov Equation

S. -T. Yau,S. S. -T. Yau

DOI: https://doi.org/10.1007/bf01182625

1996-01-01

Abstract:The original motivation of this study was to solve the Duncan-Mortensen-Zakai equation in nonlinear filtering theory. It was found that to do so one has to solve the so-called Kolmogorov equation. A scheme for solving the Kolmogorov equation explicitly is outlined.< >
An Efficient Algorithm of Yau-Yau Method for Solving Nonlinear Filtering Problems.

Mei-Heng Yueh,Wen-Wei Lin,Shing-Tung Yau

DOI: https://doi.org/10.4310/cis.2014.v14.n2.a3

2014-01-01

Communications in Information and Systems

Abstract:It is well known that the nonlinear filter has important applications in military, engineering and commercial industries. In this paper, we propose efficient and accurate numerical algorithms for the realization of the Yau-Yau method for solving nonlinear filtering problems by using finite difference schemes. The Yau-Yau method reduces the nonlinear filtering problem to the initial-value problem of Kolmogorov equations. We first solve this problem by the implicit Euler method, which is stable in most cases, but costly. Then, we propose a quasi-implicit Euler method which is feasible for acceleration by fast Fourier transformations. Furthermore, we propose a superposition technique which enables us to deal with the nonlinear filtering problem in an off-time process and thus, save a large amount of computational cost. Next, we prove that the numerical solutions of Kolmogorov equations by our schemes are always nonnegative in each iteration. Consequently, our iterative process preserves the probability density functions. In addition, we prove convergence of our schemes under some mild conditions. Numerical results show that the proposed algorithms are efficient and promising.
Linear Filtering with Nonlinear Observations

SST Yau,ST Yau,CL Yan

DOI: https://doi.org/10.1109/cdc.2004.1430360

2004-01-01

Abstract:For all known finite dimensional filters, one always needs the condition that the observation terms are degree one polynomials. On the other hand, in many practical examples, e.g. tracking problems, the observation terms may be nonlinear. In this paper, we develop a new method to treat linear filtering problems with nonlinear observation terms. We first show that real time computation of Duncan-Mortensen-Zakai (DMZ) equation can be reduced to off time computation of Kolmogorov equation. An explicit algorithm of such a reduction is provided. We next show that if the drifts are linear and the observation terms are nonlinear with linear growths, then the Kolmogorov equation can be solved in real time via a system of ODEs. Consequently, the nonlinear filtering problem with linear drifts and nonlinear observations with linear growth can be solved in real time and memoryless manner.
Hermite Spectral Method to 1-D Forward Kolmogorov Equation and Its Application to Nonlinear Filtering Problems

Xue Luo,Stephen S.–T. Yau

DOI: https://doi.org/10.1109/tac.2013.2259975

IF: 6.549

2013-01-01

IEEE Transactions on Automatic Control

Abstract:In this paper, we investigate the Hermite spectral method (HSM) to numerically solve the forward Kolmogorov equation (FKE). A useful guideline of choosing the scaling factor of the generalized Hermite functions is given in this paper. It greatly improves the resolution of HSM. The convergence rate of HSM to FKE is analyzed in the suitable function space and has been verified by the numerical simulation. As an important application and our primary motivation to study the HSM to FKE, we work on the implementation of the nonlinear filtering (NLF) problem with a real-time algorithm developed in [17]. The HSM to FKE is served as the off-line computation in this algorithm. The translating factor of the generalized Hermite functions and the moving-window technique are introduced to deal with the drifting of the posterior conditional density function of the states in the on-line experiments. Two numerical experiments of NLF problems are carried out to illustrate the feasibility of our algorithm. Moreover, our algorithm surpasses the particle filter as a real-time solver to NLF.
New Algorithms in Real Time Solution of the Nonlinear Filtering Problem.

Stephen S.-T. Yau

DOI: https://doi.org/10.4310/cis.2008.v8.n3.a6

2008-01-01

Abstract:It is well known that the filtering theory has important applications in both military and commercial industries.The Kalman-Bucy filter has been used in many areas such as navigational and guidance systems, radar tracking, solar mapping, and satellite orbit determination.However, the Kalman-Bucy filter has limited applicability because of the linearity assumptions of the drift term and observation term as well as the Gaussian assumption of the initial value.Therefore there has been an intensive interest in solving the nonlinear filtering problem.The central problem of nonlinear filtering theory is to solve the DMZ equation in real time and memoryless way.In this paper, we shall describe three methods to solve the DMZ equation: Brockett-Mitter estimation algebra method, direct method, and new algorithm method.The first two methods are relatively easy to implement in hardware and can solve a large class of nonlinear filtering problems.We shall present the recent advance in the third method which solves all the nonlinear filtering problems in a real-time manner in theory.
Proper Orthogonal Decomposition Method to Nonlinear Filtering Problems in Medium-High Dimension

Zhongjian Wang,Xue Luo,Stephen S-T Yau,Zhiwen Zhang

DOI: https://doi.org/10.1109/tac.2019.2927322

IF: 6.549

2020-01-01

IEEE Transactions on Automatic Control

Abstract:In this paper, we investigate the proper orthogonal decomposition (POD) method to numerically solve the forward Kolmogorov equation (FKE). Our method aims to explore the low-dimensional structures in the solution space of the FKE and to develop efficient numerical methods. As an important application and our primary motivation to study the POD method to FKE, we solve the nonlinear filtering (NLF) problems with a real-time algorithm proposed by Yau and Yau combined with the POD method. This algorithm is referred as POD algorithm in this paper. Our POD algorithm consists of offline and online stages. In the offline stage, we construct a small number of POD basis functions that capture the dynamics of the system and compute propagation of the POD basis functions under the FKE operator. In the online stage, we synchronize the coming observations in a real-time manner. Its convergence analysis has also been discussed. Some numerical experiments of the NLF problems are performed to illustrate the feasibility of our algorithm and to verify the convergence rate. Our numerical results show that the POD algorithm provides considerable computational savings over existing numerical methods.
Explicit Solution to Generalized Time-Varying Kolmogorov Equation

Xi Wu,Yau, S.S.-T.

DOI: https://doi.org/10.1109/cdc.1999.830912

1999-01-01

Abstract:In this paper, we give a closed formal solution to generalized Kolmogorov equation. Estimates of the formal solution are obtained. From the truncated formal solution, we are able to construct the convergent fundamental solution and estimate the time interval on which our solution converges
Direct Method for Time-Varying Nonlinear Filtering Problems

Xiuqiong Chen,Xue Luo,Stephen S. -T. Yau

DOI: https://doi.org/10.1109/taes.2017.2651650

IF: 3.491

2017-01-01

IEEE Transactions on Aerospace and Electronic Systems

Abstract:This paper discusses how to solve a filtering problem for a class of continuous nonlinear time-varying systems via the Duncan-Mortensen-Zakai (DMZ) equation. In this paper, the original DMZ equation is changed into the Kolmogorov forward equation (KFE) by exponential transformations in each time interval, and then, under some assumptions, the KFE can be transformed into a time-varying Schrodinger equation, which can be solved explicitly. The novelty of this paper lies in how to transform the KFE into the Schrodinger equation. As a direct application, the results of the paper "Nonlinear filtering and time varying Schrodinger equation" are extended for time-varying Yau systems.
Real Time Algorithm for Nonlinear Filtering Problem.

SST Yau,ST Yau

DOI: https://doi.org/10.1109/.2001.980569

2002-01-01

Abstract:The fundamental problem of nonlinear filtering theory is how to solve robust D-M-Z equation in real time and in memoryless manner. This paper describes a new real time algorithm which reduces the nonlinear filtering problem to off-line computations. Our algorithm gives convergent solutions in both pointwise sense and L/sup 2/ in case that the drift term and observation dynamic term have linear growths. The algorithm presented is slightly better than that given in our previous paper (2000).
Schr&#x00F6;dinger Equation with Quartic Potential and Nonlinear Filtering Problem

Der–Chen Chang,Stephen S.‐T. Yau,Ke-Pao Lin

DOI: https://doi.org/10.1109/cdc.2009.5400128

2009-01-01

Abstract:The celebrated work of Yau and Yau solved the nonlinear filtering problem in theory in the following manner. They reduced the problem of solving the Duncan-Mortensen-Zakai equation in real-time to the off-time solution of a Kolmogorov type equation. For the Yau filtering system, this Kolmogorov equation can be transformed as the Schrodinger equation. In this paper, we shall describe the fundamental solution of this Schrodinger equation with quartic potential.
Parallel Computations for Yau Filters

HW Cheng,SST Yau

DOI: https://doi.org/10.1016/j.camwa.2005.01.026

IF: 3.218

2005-01-01

Computers & Mathematics with Applications

Abstract:In the early 1990s, Yau developed a new class of nonlinear filters, called the Yau Filterswhich contain the Kalman-Bucy filters and the Benes filters as special cases. It has been shown that, from the Lie algebraic point of view, the Yau filters are the most general finite-dimensional filters. Yau and Hu proved that the DMZ equation for a Yau filter can be reduced to a Kolmogorov type partial differential equation and a system of linear differential equations. They noticed that the PDE is independent of the observed data and hence can be solved off-line. An efficient parallel algorithm for the system of ODEs would lead to fast solutions of Yau filters and hence their suitability for real-life applications. In this paper, we have proposed several parallel methods suitable for this system of ODEs.
Solving Non-linear Kolmogorov Equations in Large Dimensions by Using Deep Learning: A Numerical Comparison of Discretization Schemes

Raffaele Marino,Nicolas Macris

DOI: https://doi.org/10.1007/s10915-022-02044-x

2022-11-25

Journal of Scientific Computing

Abstract:Non-linear partial differential Kolmogorov equations are successfully used to describe a wide range of time dependent phenomena, in natural sciences, engineering or even finance. For example, in physical systems, the Allen–Cahn equation describes pattern formation associated to phase transitions. In finance, instead, the Black–Scholes equation describes the evolution of the price of derivative investment instruments. Such modern applications often require to solve these equations in high-dimensional regimes in which classical approaches are ineffective. Recently, an interesting new approach based on deep learning has been introduced byby E, Han and Jentzen [1, 2]. The main idea is to construct a deep network which is trained from the samples of discrete stochastic differential equations underlying Kolmogorov's equation. The network is able to approximate, numerically at least, the solutions of the Kolmogorov equation with polynomial complexity in whole spatial domains. In this contribution we study variants of the deep networks by using different discretizations schemes of the stochastic differential equation. We compare the performance of the associated networks, on benchmarked examples, and show that, for some discretization schemes, improvements in the accuracy are possible without affecting the observed computational complexity.

mathematics, applied
A Novel Real-Time Filtering Method to General Nonlinear Filtering Problem Without Memory

Ji Shi,Xiuqiong Chen,Stephen Shing-Toung Yau

DOI: https://doi.org/10.1109/access.2021.3050572

IF: 3.9

2021-01-01

IEEE Access

Abstract:In this paper, the filtering problem for the general time-invariant nonlinear state-observation system is considered. Our work is based on the Yau-Yau filtering framework developed by S.-T. Yau and the third author in 2008. The key problem of Yau-Yau filtering framework is how to compute the solution to forward Kolmogorov equation (FKE) off-line effectively. Motivated by the supervised learning in machine learning, we develop an efficient method to numerically solve the FKE off-line from the point of view of optimization. Specifically, for the off-line computation part, the computation of the solution to a FKE is reduced to computing a linear system of equations by making the temporal inverse transformation and the loss function optimization, and we store the results for the preparation of on-line computation. For the on-line computation part, the unnormalized density function is approximated by a complete polynomial basis, and then the estimation of the state is computed using the stored off-line data. Our method has the merits of easily implementing, real-time and memoryless. More importantly, it can be applicable for moderate-high dimensional cases. Numerical experiments have been carried out to verify the feasibility of our method. Our algorithm outperforms extended Kalman filter, unscented Kalman filter and particle filter both in accuracy and costing time.
Real Time Solution of the Nonlinear Filtering Problem Without Memory Ii

ST Yau,SST Yau

DOI: https://doi.org/10.1137/050648353

IF: 2.2

2008-01-01

SIAM Journal on Control and Optimization

Abstract:It is well known that the nonlinear filtering problem has important applications in both military and commercial industries. The central problem of nonlinear filtering is to solve the Duncan-Mortensen-Zakai ( DMZ) equation in real time and in a memoryless manner. The purpose of this paper is to show that, under very mild conditions ( which essentially say that the growth of the observation vertical bar h vertical bar is greater than the growth of the drift vertical bar f vertical bar), the DMZ equation admits a unique nonnegative weak solution u which can be approximated by a solution u(R) of the DMZ equation on the ball B-R with u(R)vertical bar(partial derivative BR) = 0. The error of this approximation is bounded by a function of R which tends to zero as R goes to infinity. The solution uR can in turn be approximated efficiently by an algorithm depending only on solving the observation-independent Kolmogorov equation on BR. In theory, our algorithm can solve basically all engineering problems in real time. Specifically, we show that the solution obtained from our algorithms converges to the solution of the DMZ equation in the L-1 sense. Equally important, we have a precise error estimate of this convergence, which is important in numerical computation.
Solving Nonlinear Filtering Problems in Real Time by Legendre Galerkin Spectral Method.

Wenhui Dong,Xue Luo,Stephen S-T Yau

DOI: https://doi.org/10.1109/tac.2020.3002979

IF: 6.549

2021-01-01

IEEE Transactions on Automatic Control

Abstract:It is well known that the nonlinear filtering (NLF) problem has important applications in both military and civil industries. The central question is to solve the posterior conditional density function of the states, which satisfies the Kushner or the Duncan-Mortensen-Zakai (DMZ) equation after suitable change of probability measure. In this article, we shall follow the so-called Yau-Yau's algorithm to split the solution of the DMZ equation into on- and off-line part, where the off-line part is to solve the forward Kolmogorov equation (FKE) with the initial conditions to be the orthonormal bases in some suitable function space. Instead of the generalized Hermite function investigated by the second and the third author of this article, we shall explore the generalized Legendre polynomials. The Legendre spectral method (LSM) is used to numerically solve the FKE. Under certain conditions, the convergence rate of LSM is twice faster than that of the Hermite spectral method. Two two-dimensional numerical experiments of NLF problems (time-invariant and time-varying cases) have been numerically solved to illustrate the feasibility of our algorithm. Our algorithm outperforms the extended Kalman Filter and particle filter in both real-time manner and accuracy.
Nonlinear Filtering and Time Varying Schrodinger Equation I

ST Yau,SST Yau

DOI: https://doi.org/10.1109/taes.2004.1292160

IF: 3.491

2004-01-01

IEEE Transactions on Aerospace and Electronic Systems

Abstract:Based on our previous work we have successfully reduced the nonlinear filtering problem for Yau filtering system to the time-varying Schrodinger equation. In order to solve the nonlinear filtering problem, one needs to solve the time-varying Schrodinger equation with an arbitrary initial condition. We then solve the time-varying Schrodinger equation by constructing the fundamental solution explicitly via a system of nonlinear ODES in case the potential is quadratic in state variables. This system of nonlinear ODES is solved explicitly by the power series method.
On the Convergence Analysis of Yau-Yau Nonlinear Filtering Algorithm: from a Probabilistic Perspective

Zeju Sun,Xiuqiong Chen,Stephen S.-T. Yau

2024-05-10

Abstract:At the beginning of this century, a real time solution of the nonlinear filtering problem without memory was proposed in [1, 2] by the third author and his collaborator, and it is later on referred to as Yau-Yau algorithm. During the last two decades, a great many nonlinear filtering algorithms have been put forward and studied based on this framework. In this paper, we will generalize the results in the original works and conduct a novel convergence analysis of Yau-Yau algorithm from a probabilistic perspective. Instead of considering a particular trajectory, we estimate the expectation of the approximation error, and show that commonly-used statistics of the conditional distribution (such as conditional mean and covariance matrix) can be accurately approximated with arbitrary precision by Yau-Yau algorithm, for general nonlinear filtering systems with very liberal assumptions. This novel probabilistic version of convergence analysis is more compatible with the development of modern stochastic control theory, and will provide a more valuable theoretical guidance for practical implementations of Yau-Yau algorithm.

Optimization and Control,Probability

Recent Advance In Computing Kolmogorov Equation Arising From Nonlinear Filtering

Special Solutions to Some Kolmogorov Equations Arising from Cubic Sensor Problems.

Finite-dimensional Filters with Nonlinear Drift. V: Solution to Kolmogorov Equation Arising from Linear Filtering with Non-Gaussian Initial Condition

Wavelet-Galerkin Method for the Kolmogorov Equation

Explicit Solution of A Kolmogorov Equation

An Efficient Algorithm of Yau-Yau Method for Solving Nonlinear Filtering Problems.

Linear Filtering with Nonlinear Observations

Hermite Spectral Method to 1-D Forward Kolmogorov Equation and Its Application to Nonlinear Filtering Problems

New Algorithms in Real Time Solution of the Nonlinear Filtering Problem.

Proper Orthogonal Decomposition Method to Nonlinear Filtering Problems in Medium-High Dimension

Explicit Solution to Generalized Time-Varying Kolmogorov Equation

Direct Method for Time-Varying Nonlinear Filtering Problems

Real Time Algorithm for Nonlinear Filtering Problem.

Schr&amp;#x00F6;dinger Equation with Quartic Potential and Nonlinear Filtering Problem

Parallel Computations for Yau Filters

Solving Non-linear Kolmogorov Equations in Large Dimensions by Using Deep Learning: A Numerical Comparison of Discretization Schemes

A Novel Real-Time Filtering Method to General Nonlinear Filtering Problem Without Memory

Real Time Solution of the Nonlinear Filtering Problem Without Memory Ii

Solving Nonlinear Filtering Problems in Real Time by Legendre Galerkin Spectral Method.

Nonlinear Filtering and Time Varying Schrodinger Equation I

On the Convergence Analysis of Yau-Yau Nonlinear Filtering Algorithm: from a Probabilistic Perspective

Schrödinger Equation with Quartic Potential and Nonlinear Filtering Problem