解析虚拟现实抓取难题，推动沉浸式交互体验提升
Grasping Object: Challenges and Innovations in Robotics and Virtual Reality
- 梳理虚拟现实与机器人抓取技术的异同与挑战
- 指出当前方案在视觉与触觉反馈上难以兼顾真实感
- 适合关注虚拟现实交互、人机协同的研究者参考
现实世界中，抓取是操作物体的基本且高效的方式，这一原则同样适用于物理与虚拟世界。然而，虚拟现实中抓取涉及图形、物理模拟与感知的复杂结合，用户沉浸感依赖于逼真的触觉反馈与高质量画面，而这些在实时场景中计算成本高且难以实现。现有方法在模拟不同动态物体的抓取时，无法同时提供可信的视觉效果与触觉反馈。本文综述了虚拟现实与机器人领域中的抓取技术，揭示其核心挑战，探讨手部抓取不同动力学物体的复杂性，并激发针对虚拟现实应用的创新解决方案。
原文摘要 · Abstract (English)
In real life, grasping is one of the fundamental and effective forms of interaction when manipulating objects. This holds true in the physical and virtual world; however, unlike the physical world, virtual reality (VR) is grasped in a complex formulation that includes graphics, physics, and perception. In virtual reality, the user's immersion level depends on realistic haptic feedback and high-quality graphics, which are computationally demanding and hard to achieve in real-time. Current solutions fail to produce plausible visuals and haptic feedback when simulation grasping in VR with a variety of targeted object dynamics. In this paper, we review the existing techniques for grasping in VR and robotics and indicate the main challenges that grasping faces in the domains. We aim to explore and understand the complexity of hand-grasping objects with different dynamics and inspire various ideas to improve and come up with potential solutions suitable for virtual reality applications.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。