3 accepted papers
Learning agile humanoid behaviors from human motion offers a powerful route to natural, coordinated control, but existing approaches face a persistent trade-off: reference-tracking policies are often brittle outside the demonstration dataset, while purely task-driven Reinforcement Learning (RL) can …
Differentiable simulators promise to improve sample efficiency in robot learning by providing analytic gradients of the system dynamics. Yet, their application to contact-rich tasks like locomotion is complicated by the inherently non-smooth nature of contact, impeding effective gradient-based optim…
Reinforcement learning (RL) has shown remarkable proficiency in developing robust control policies for contact-rich applications. However, it typically requires meticulous Markov Decision Process (MDP) designing tailored to each task and robotic platform. This work addresses this challenge by creati…