The field of human-robot interaction and manipulation is rapidly evolving, with a focus on developing more intuitive and efficient interfaces. Researchers are exploring the use of embodied gesture-driven control, woven tactile skins, and affordance-aware approaches to improve human-robot collaboration. Notable advancements include the development of novel computational design pipelines, sensing techniques, and diffusion-based generative models. These innovations have the potential to enhance the accuracy and reliability of robotic systems, particularly in tasks such as food scooping, powder manipulation, and 3D object reconstruction.

Some noteworthy papers in this area include: DHAGrasp, which proposes a text-guided dual-hand grasp generator that synthesizes affordance-aware grasps for unseen objects. WireBend-kit, which introduces a computational design and fabrication toolkit for creating custom 3D wireframe structures. Affordance-Guided Diffusion Prior, which presents a generative prior for hand pose refinement guided by affordance-aware textual descriptions of hand-object interactions. GRITS, which leverages guided diffusion policy to minimize food spillage during scooping and ensure reliable transfer of food items. Differentiable Skill Optimisation, which proposes a trajectory optimisation framework for powder transport in laboratory settings.