7 papers
LyEvO: Lyapunov-Guided Evolutionary Optimization for Safe and Robust Sim-to-Real Policy Learning
Riccardo Curcio, Hongpeng Cao, Marco Caccamo
Training controllers that are safe and robust in simulation, and systematically assessing their readiness for real-world deployment, remain key challenges in sim-to-real transfer.…
Safe Online Learning via Smooth Safety-Structured Policy Composition
Hongpeng Cao, Liqun Zhao, Yuliang Gu +3
Safe online reinforcement learning requires policies to respect safety constraints while maintaining smooth optimization dynamics. Existing approaches typically rely on either stri…
Runtime Learning of Quadruped Robots in Wild Environments
Yihao Cai, Yanbing Mao, Lui Sha +2
This paper presents a runtime learning framework for quadruped robots, enabling them to learn and adapt safely in dynamic wild environments. The framework integrates sensing, navig…
Observations Meet Actions: Learning Control-Sufficient Representations for Robust Policy Generalization
Yuliang Gu, Hongpeng Cao, Marco Caccamo +1
Capturing latent variations ("contexts") is key to deploying reinforcement-learning (RL) agents beyond their training regime. We recast context-based RL as a dual inference-control…
Bregman Centroid Guided Cross-Entropy Method
Yuliang Gu, Hongpeng Cao, Marco Caccamo +1
The Cross-Entropy Method (CEM) is a widely adopted trajectory optimizer in model-based reinforcement learning (MBRL), but its unimodal sampling strategy often leads to premature co…
Physics-model-guided Worst-case Sampling for Safe Reinforcement Learning
Hongpeng Cao, Yanbing Mao, Lui Sha +1
Real-world accidents in learning-enabled CPS frequently occur in challenging corner cases. During the training of deep reinforcement learning (DRL) policy, the standard setup for t…