Estimator Design

The ability to infer the properties of a system from only incomplete or partial information is commonly referred to as an estimation problem. For application to fluid mechanics, a typical situation is to attempt to infer flow-field properties from boundary data. Two significant challenges are that fluid flows of practical interest are nonlinear and evolve over very short timescales, meaning that the flow estimates must be obtained at high frequency. Moving Horizon Estimation is an optimization-based methodology that can address these two difficulties and we develop theoretical properties of such optimization schemes.

Schematic of a one-step MHE scheme

Multiple-shooting and single-shooting MHE for the Lorenz attractor

Related Publications

Ji L, Rawlings JB, Hu W, Wynn A, Diehl M, 2015, Robust Stability of Moving Horizon Estimation Under Bounded Disturbances, IEEE Transactions on Automatic Control, Pages: 1-1, Link
Wynn A., Vukov M., Diehl M., 2014, Convergence guarantees for moving horizon estimation based on the real-time iteration scheme, IEEE Transactions on Automatic Control, 59, 2215-2221. Link.
Wynn A., Goulart P.J., 2011, Observer design for systems with an energy preserving nonlinearity, with application to fluid flow, 50th IEEE Conference of Decision and Control / European Control Conference, 7524-7529. Link.