RCUKF: Data-Driven Modeling Meets Bayesian Estimation

TL;DR

The paper introduces RCUKF, a hybrid framework combining reservoir computing and unscented Kalman filtering to improve data-driven modeling and state estimation in complex systems, especially where traditional models struggle.

Contribution

It presents a novel integration of reservoir computing with UKF, enabling real-time, accurate state estimation using learned nonlinear dynamics from data.

Findings

Effective in high-dimensional and chaotic regimes

Improves state estimation accuracy with real-time sensor data

Validated on benchmark problems and vehicle trajectory tasks

Abstract

Accurate modeling is crucial in many engineering and scientific applications, yet obtaining a reliable process model for complex systems is often challenging. To address this challenge, we propose a novel framework, reservoir computing with unscented Kalman filtering (RCUKF), which integrates data-driven modeling via reservoir computing (RC) with Bayesian estimation through the unscented Kalman filter (UKF). The RC component learns the nonlinear system dynamics directly from data, serving as a surrogate process model in the UKF prediction step to generate state estimates in high-dimensional or chaotic regimes where nominal mathematical models may fail. Meanwhile, the UKF measurement update integrates real-time sensor data to correct potential drift in the data-driven model. We demonstrate RCUKF effectiveness on well-known benchmark problems and a real-time vehicle trajectory estimation task in a high-fidelity simulation environment.