Derives an ε-local minimax excess-cost lower bound for learning LQG controllers from offline trajectories of a linear exploration policy, expressed via the Hessian of the LQG cost and inverse Fisher information, and instantiates it on fragile robust-control examples.
Certainty equivalence is efficient for linear quadratic control
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
verdicts
UNVERDICTED 3representative citing papers
Relearn LQR combines recursive least squares with policy gradient for on-policy data-driven LQR and proves stability of the full scheme via Lyapunov analysis with averaging and timescale separation.
A new regularized covariance parameterization enables effective direct data-driven LQR control for ill-conditioned data, shown equivalent to indirect Tikhonov-regularized LQR and extended to nonlinear systems via Koopman embedding.
citing papers explorer
-
The Fragility of Learning LQG Controllers
Derives an ε-local minimax excess-cost lower bound for learning LQG controllers from offline trajectories of a linear exploration policy, expressed via the Hessian of the LQG cost and inverse Fisher information, and instantiates it on fragile robust-control examples.
-
Stability-Certified On-Policy Data-Driven LQR via Recursive Learning and Policy Gradient
Relearn LQR combines recursive least squares with policy gradient for on-policy data-driven LQR and proves stability of the full scheme via Lyapunov analysis with averaging and timescale separation.