Non-conventional Estimation Theorems Concerning a Ubiquitous Bilinear Stochastic Differential System: a Control Perspective

TL;DR

This paper develops a formal estimation theory for vector Stratonovich bilinear stochastic differential equations, with applications to physical systems like a stochastic three-phase rectifier circuit, advancing control and estimation methods.

Contribution

It introduces a systematic estimation framework for non-homogeneous Markov processes governed by vector Stratonovich bilinear SDEs, filling a gap in the existing theory.

Findings

Derived estimation theorems for bilinear stochastic systems

Applied the theory to a stochastic three-phase rectifier circuit

Enhanced control and estimation techniques for nonlinear physical systems

Abstract

Stochastic differential equations and the associated partial differential equations are the cornerstone formalism in stochastic control problems. The universality of bilinear stochastic systems can be found in autonomous systems, non-linear dynamic circuits, and mathematical finance. Consensus on the Ito versus Stratonovich suggests stochastic systems embedded with Stratonovich differential to describe the stochastic evolution of the state vector of real physical systems. The mathematical theory of a scalar time-varying bilinear Stratonovich stochastic differential equation is available in current texts. The theory of scalar Stratonovich systems was developed by deriving their closed-form solutions and then conditional moments. Practical problems obeying vector time-varying bilinear Stratonovich stochastic differential equations are ubiquitous. However, their formal and systematic estimation theory is not available. In this paper, we develop the results for non-homogeneous Markov processes obeying the vector Stratonovich bilinear stochastic differential equation. Then, the theory of the paper is applied to a stochastic three-phase rectifier circuit. The stochastic evolution of the rectifier state vector is constructed by utilizing the Euler-Lagrange theory and embedding Stratonovich differential. Since the theory of the paper accounts for the multidimensionality as well as respects Stratonovich stochasticity of the ubiquitous bilinear system, the estimation Theorems will be remarkably useful to estimation and control of nonlinear real physical systems. Finally, this paper will be of interest to control and computational practitioners aspiring for the formal theory of bilinear stochastic systems arising from their application as well as applied mathematicians looking for applications of formal bilinear stochastic estimation theory.