Control and Estimation Co-Design via Envelope-Theorem Gradients
Abstract
Instead of the sequential plant--control--estimator design pipeline, which is in general not optimal, we propose the Control and Estimation Co-Design (ContEst) framework, which poses the entire system design as a single two-stage problem solved by first-order methods. A key enabler is that the gradient of each inner control/estimation optimal cost with respect to the design parameters is available from the dual variables the inner solver already returns, by the envelope theorem: exactly under mild regularity (a unique inner minimizer), and as a subgradient under more relaxed conditions. That is, no differentiation through the optimizer or its optimality conditions is needed. We first present the framework in a linear quadratic Gaussian (LQG) regime, where the control and estimation costs become two semidefinite programs (SDPs) whose design gradients are read from the duals of the Lyapunov constraints. We extend the results to robust (H∞) formulation to cover worst-case control and filtering problems. We also present an extension to constrained and nonlinear systems, using an extended Kalman filter (eKF) and a local linearization of the information-state dynamics to yield an approximate but convex model predictive control (MPC) problem whose design gradients are read directly from the costates. We apply our methods to a variety of realistic design examples from different fields and report significant design improvements.
Related
- A semi-parametric Bernstein-von Mises theorem at rate n−3/2 for frequency estimation
- Minimax rates for pointwise CATE estimation: the role of design smoothness
- Nearly-minimax variance estimation under rough random design
- Spherical-Design Oracles for Zeroth-Order Optimisation: Deterministic, Variance-Free Gradient Estimation
- Trustworthy Method Comparison with AI Judges: Estimation and Design under Order, Batch, and Aggregation Effects
BibTeX
@misc{ramadan2026control,
title = {{Control and Estimation Co-Design via Envelope-Theorem Gradients}},
author = {Mohammad S. Ramadan and Philip Dinenis and Mihai Anitescu},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.36090},
eprint = {2609.36090},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.36090},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Control and Estimation Co-Design via Envelope-Theorem Gradients
AU - Mohammad S. Ramadan
AU - Philip Dinenis
AU - Mihai Anitescu
PY - 2026
DA - 2026/09/28
PB - arXiv
DO - 10.48550/arxiv.2609.36090
UR - https://arxiv.org/abs/2609.36090
AB - Instead of the sequential plant--control--estimator design pipeline, which is in general not optimal, we propose the Control and Estimation Co-Design (ContEst) framework, which poses the entire system design as a single two-stage problem solved by first-order methods. A key enabler is that the gradient of each inner control/estimation optimal cost with respect to the design parameters is available from the dual variables the inner solver already returns, by the envelope theorem: exactly under mild regularity (a unique inner minimizer), and as a subgradient under more relaxed conditions. That is, no differentiation through the optimizer or its optimality conditions is needed. We first present the framework in a linear quadratic Gaussian (LQG) regime, where the control and estimation costs become two semidefinite programs (SDPs) whose design gradients are read from the duals of the Lyapunov constraints. We extend the results to robust (H_∞) formulation to cover worst-case control and filtering problems. We also present an extension to constrained and nonlinear systems, using an extended Kalman filter (eKF) and a local linearization of the information-state dynamics to yield an approximate but convex model predictive control (MPC) problem whose design gradients are read directly from the costates. We apply our methods to a variety of realistic design examples from different fields and report significant design improvements.
ER -