Reinforced Optimal Estimator

Cao, Wenhan; Chen, Jianyu; Duan, Jingliang; Li, Shengbo Eben; Lyu, Yao; Gu, Ziqing; Zhang, Yuhang

doi:10.1016/j.ifacol.2021.11.201

Cited by 2 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Apply a backup estimator to obtain the estimate V. STABILITY ANALYSIS In this section, we prove the stability of the offline primal estimator solved by (15). Our analysis follows similar ideas in [28], with suitable extensions to account for the "∆suboptimality" of the learned estimator.…”

Section: Repeatmentioning

confidence: 99%

“…Offline Training 1: Learn primal estimator X(I; η * ) as in (15) 2: Learn dual estimator Λ(I; ω * ) as in (16) 3: Validate X(I; η * ) and Λ(I; ω * ) using Theorem 1…”

Section: Algorithm 1 Primal-dual Mhementioning

confidence: 99%

“…In particular, one can parameterize the MHE estimator using neural networks or a linear combination of basis functions, and then find a parameterized estimator that minimizes the MHE cost using supervised learning techniques. Some studies also apply reinforcement learning or variational inference to obtain an offline estimator [15]- [18]. However, existing offline estimation methods lack the ability to verify the estimation accuracy in real-time during the online application process.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Primal-dual Estimator Learning: an Offline Constrained Moving Horizon Estimation Method with Feasibility and Near-optimality Guarantees

Cao¹,

Duan²,

Li³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

This paper proposes a primal-dual framework to learn a stable estimator for linear constrained estimation problems leveraging the moving horizon approach. To avoid the online computational burden in most existing methods, we learn a parameterized function offline to approximate the primal estimate. Meanwhile, a dual estimator is trained to check the suboptimality of the primal estimator during execution time. Both the primal and dual estimators are learned from data using supervised learning techniques, and the explicit sample size is provided, which enables us to guarantee the quality of each learned estimator in terms of feasibility and optimality. This in turn allows us to bound the probability of the learned estimator being infeasible or suboptimal. Furthermore, we analyze the stability of the resulting estimator with a bounded error in the minimization of the cost function. Since our algorithm does not require the solution of an optimization problem during runtime, state estimates can be generated online almost instantly. Simulation results are presented to show the accuracy and time efficiency of the proposed framework compared to online optimization of moving horizon estimation and Kalman filter. To the best of our knowledge, this is the first learning-based state estimator with feasibility and nearoptimality guarantees for linear constrained systems.

show abstract

Section: Repeatmentioning

confidence: 99%

“…Offline Training 1: Learn primal estimator X(I; η * ) as in (15) 2: Learn dual estimator Λ(I; ω * ) as in (16) 3: Validate X(I; η * ) and Λ(I; ω * ) using Theorem 1…”

Section: Algorithm 1 Primal-dual Mhementioning

confidence: 99%