|
Humanoid robotics has advanced rapidly in recent years. While research primarily focused on agile motion using model-based whole-body control and reinforcement learning, the true advantage of humanoids lies in their human-like morphology. This allows them to perform tasks autonomously in environments designed for humans, making them uniquely suited for supporting human activities. To fully realize this potential, dexterous manipulation is essential. Although hardware design, control
strategies, and high-level decision-making have each advanced substantially within their respective domains, these components remain inherently interdependent. For example, robust hardware and control algorithms are essential for effective autonomy, while high-level decision-making requirements often drive hardware and controller design. Nevertheless, collaboration across these areas remains limited, largely because of divergent technical focuses and expertise.
Therefore, this workshop aims to bridge these communities by exploring synergies across robot hardware, control algorithms, and high-level autonomy, with an emphasis on enabling dexterous manipulation for humanoid upper bodies. We are particularly interested in hardware designs that support versatile manipulation, control strategies for stable and contact-rich interaction, and learning-based frameworks for developing cost-effective autonomous systems. Topics of discussion will include dexterous hand design, tactile sensing, whole-body planning and control, and learning approaches for skill acquisition and deployment. We invite researchers from both academia and industry to contribute their perspectives. By bringing together diverse expertise, we aim to catalyze collaboration and accelerate progress toward autonomous, dexterous humanoid systems.
|
