Abstract
Physical objects serving as haptic proxies offer a promising approach to enrich tactile experience in virtual reality. However, conventional haptic proxies are hampered by prohibitive costs, limited reusability, and the logistical burden of creating and storing numerous object-specific models. Here, we introduce a fabric-based topological haptic proxy (FTHP) that functions as a programmable and universal interface. Our approach synergistically integrates origami-inspired topological constraints with triboelectric sensor yarns. The engineered topological design, featuring heterogeneous rigid and flexible segments, restricts deformation pathways, ensuring structural stability and generating distinct, classifiable electrical signals for different interactions. This allows a single, reusable FTHP to be dynamically reconfigured into multiple functional states (e.g. a flat touchpad or various 3D geometric controllers), bypassing the need for a rigid one-to-one correspondence between physical props and pre-stored virtual assets. Integrated with a convolutional neural network (CNN), the system achieves a 92.4% recognition accuracy across 14 distinct actions and 3 interaction modes. The FTHP presents a scalable and versatile platform for high-fidelity haptic interaction, advancing the design of more immersive and accessible virtual reality (VR) systems.