Joint system identification and state estimation for diverse sets of targets - Online methods for SDE state-space models with latent states

Research Project, 2025 – 2029

This project studies methods for combining online system identification with state estimation in stochastic differential equation (SDE) models using streaming data. The goal is to use incoming observations to update estimates of a system’s hidden state while also refining unknown, potentially time-varying parameters or aspects of the governing dynamics when the data supports it. This type of framework is relevant in settings where both uncertainty and adaptation matter, including robotics and autonomous systems, finance, climate and geophysical monitoring, and biomedical or engineering applications with real-time sensors.