Abstract:The work of Kalman and Bucy has established a duality between filtering and optimal estimation in the context of time-continuous linear systems. This duality has recently been extended to time-continuous nonlinear systems in terms of an optimization problem constrained by a backward stochastic partial differential equation. Here we revisit this problem from the perspective of appropriate forward-backward stochastic differential equations. This approach sheds new light on the estimation problem and provides a unifying perspective. It is also demonstrated that certain formulations of the estimation problem lead to deterministic formulations similar to the linear Gaussian case as originally investigated by Kalman and Bucy. Finally, optimal control of partially observed diffusion processes is discussed as an application of the proposed estimators.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the forward - backward stochastic differential equation (FBSDE) method in continuous - time minimum variance estimation. Specifically, the paper aims to re - examine and extend the duality between filtering and optimal estimation in linear systems established by Kalman and Bucy, generalize it to nonlinear systems, and provide a unified perspective by introducing appropriate forward - backward stochastic differential equations. The main contributions of the paper include: 1. **Re - examining the estimation problem**: Re - examine the estimation problem in continuous - time filtering problems from the perspective of FBSDE, especially for the treatment of nonlinear systems. 2. **Unified framework**: Propose an FBSDE - based method that can uniformly handle the following four cases: - (i) Estimate \(\pi_T[f]\) using an estimator of the form \(S_T[f]=\mu[Y_0]-\int_{0}^{T}U_t^{T}dZ_t\); - (ii) Estimate \(\sigma_T[f]\) using an estimator of the same form, where \(\sigma_t\) is the unnormalized filtering distribution satisfying the Zakai equation; - (iii) Estimate \(\pi_T[f]\) by using the innovation process \(I_t = Z_t-\int_{0}^{t}\pi_s[h]ds\) instead of the observation value \(Z_t\); - (iv) Estimate \(\sigma_T[f]\) by using the induced observation error \(W_t = Z_t-\int_{0}^{t}h(X_s)ds\) instead of the direct observation value \(Z_t\). 3. **Deterministic equations**: Discover that the estimation problems in cases (ii) - (iv) can lead to deterministic backward Kolmogorov equations, rather than the backward stochastic partial differential equation (BSPDE) in case (i). 4. **Numerical treatment**: By strictly following the FBSDE method, numerical treatment becomes possible, which provides a basis for future computational methods. 5. **Application expansion**: Discuss the application of the optimal control problem of partially observable diffusion processes and briefly introduce the more general conditional expectation value estimation problem. In summary, by introducing the FBSDE method, this paper not only provides a new perspective for continuous - time minimum variance estimation, but also provides a unified and effective framework for dealing with filtering and optimal estimation problems in nonlinear systems.

On forward-backward SDE approaches to continuous-time minimum variance estimation

State Estimation For Nonlinear Systems By Using Dynamic Optimization

State Estimation for Continuous-Discrete-Time Nonlinear Stochastic Systems

Continuous-time stochastic gradient descent for optimizing over the stationary distribution of stochastic differential equations

Suboptimal Linear Estimation for Continuous–discrete Bilinear Systems

Time-inconsistent Linear Quadratic Optimal Control Problem for Forward-Backward Stochastic Differential Equations

Time-inconsistent linear quadratic optimal control problem for forward–backward stochastic differential equations

Optimal control for parameter estimation in partially observed hypoelliptic stochastic differential equations

System Identification for Continuous-time Linear Dynamical Systems

Controlled Reflected SDEs and Neumann Problem for Backward SPDEs

Kalman Filter and its Modern Extensions for the Continuous-time Nonlinear Filtering Problem

Application of Kalman Filter in Stochastic Differential Equations

Continuous time Stochastic optimal control under discrete time partial observations

Duality for nonlinear filtering

Backward Stochastic Differential Equations with Conditional Reflection and Related Recursive Optimal Control Problems

Bounding the Estimation Error Covariance for Nonlinear Systems

A Concentration-Based Approach for Optimizing the Estimation Performance in Stochastic Sensor Selection

Continuity of Filters for Discrete-Time Control Problems Defined by Explicit Equations

A continuous time formulation of stochastic dual control to avoid the curse of dimensionality

Forward and Backward Mean-Field Stochastic Partial Differential Equation and Optimal Control

On derivative-free extended Kalman filtering and its Matlab-oriented square-root implementations for state estimation in continuous-discrete nonlinear stochastic systems