Neural Operators for Delay-Compensating Control of Hyperbolic PIDEs

1 code implementation • 21 Jul 2023 • Jie Qi, Jing Zhang, Miroslav Krstic

The recently introduced DeepONet operator-learning framework for PDE control is extended from the results for basic hyperbolic and parabolic PDEs to an advanced hyperbolic class that involves delays on both the state and the system output or input.

Operator learning

Proximal Policy Optimization Learning based Control of Congested Freeway Traffic

no code implementations • 12 Apr 2022 • Shurong Mo, Nailong Wu, Jie Qi, Anqi Pan, Zhiguang Feng, Huaicheng Yan, Yueying Wang

The control gains for the three feedbacks are learned from the interaction between the PPO and the numerical simulator of the traffic system without knowing the system dynamics.

Delay-Compensated Distributed PDE Control of Traffic with Connected/Automated Vehicles

no code implementations • 19 Jul 2021 • Jie Qi, Shurong Mo, Miroslav Krstic

For the linearized system, a novel three-branch bakcstepping transformation with explicit kernel functions is introduced to compensate the input delay.