A GAI-Based Haptic Transmission Architecture for Extending Headset Lifespan in Haptic-Enhanced XR

Moving computing components from headsets to cloud servers is a promising approach to increasing headsets' lifespan and comfort for Extended Reality (XR) users. However, latency is an unavoidable challenge for XR services due to longdistance transmission, especially when haptic feedback (usually requires 1 ms latency) is involved. To address this challenge, we leverage the Latent Diffusion Model (LDM) to propose a novel transmission architecture, which can accurately generate the future potential haptic feedback from current or previous video frames. Moreover, we also introduce an acceleration architecture to accelerate the haptic feedback generation process. The simulations indicate that the lifespan of headsets can be tripled by moving computing resources to the cloud. In addition, our proposed architecture can accurately generate future potential haptic feedback at least 170 ms before contact, which can satisfy the 1 ms latency requirement of haptic feedback.