arXiv:2606. 15594v1 Announce Type: cross Abstract: We present SLS^2, a framework for safe feedback motion planning from pixels using robust model predictive control (MPC) in learned latent world models.
Model predictive control is a high-performance control method, but it incurs comparatively high online computational costs across many control tasks and is challenging to realize in low-level hardware...
Reinforcement learning (RL) enables the synthesis of control policies directly from data, making it highly appealing for complex cyber-physical systems (CPSs) and robotics. A persistent challenge, however, is ensuring strict, hard safety constraints during the active learning phase.
The paper introduces A‑SLS, a learning‑aware robust adaptive model predictive control framework that forecasts future parameter uncertainty and preserves correlations among disturbances, parameter errors, and estimate updates to build tight reachable sets. It presents a GPU‑accelerated solver enabling real‑time performance on high‑dimensional systems, demonstrated on a quadruped with 61 states and a Crazyflie drone that learns dynamics online to stabilize during flight.
arXiv:2511. 05355v3 Announce Type: replace Abstract: Flow matching (FM) has shown promising results in data-driven planning.
arXiv:2606. 14536v1 Announce Type: new Abstract: Safe reinforcement learning (RL) aims to learn policies that optimize rewards while satisfying constraints.
arXiv:2508. 19186v2 Announce Type: replace-cross Abstract: Reactive obstacle avoidance methods often cause agents to become trapped in local minima, because they can often only reason one step ahead (i.
arXiv:2606. 30935v1 Announce Type: cross Abstract: While neural network control policies are powerful, their deployment on safety critical systems depends on ensuring that they obey strict constraints.
arXiv:2608. 10332v1 Announce Type: cross Abstract: Differentiable predictive control (DPC), a self-supervised learning approach for approximating explicit model predictive control (MPC) policies, offers significant computational advantages over online optimization-based MPC.
The paper presents a framework for creating neural network controllers that inherit the safety and stability guarantees of a robust model predictive control (MPC) policy. By treating the neural network’s approximation error as a bounded input disturbance and incorporating it into the MPC design, the method ensures recursive feasibility and stability before deployment. The approach is offline and constructive, requiring only Lipschitz bounds, and simulations show the learned controller matches the original MPC’s closed‑loop behavior while reducing online computation.
arXiv:2607. 07252v1 Announce Type: new Abstract: Reinforcement learning (RL) enables the synthesis of control policies directly from data, making it highly appealing for complex cyber-physical systems (CPSs) and robotics.
arXiv:2608. 03662v1 Announce Type: new Abstract: Safety shields are runtime enforcement mechanisms that restrict the actions of a controller to guarantee safety.