VkVIO: First cross-platform GPU-accelerated VIO using vendor-agnostic Vulkan
zhenjun_zhao · x · 2026-09-29
Ole Hoffmann, Mateo de Mayo, and Daniel Cremers present VkVIO, reportedly the first cross-platform GPU-accelerated visual-inertial odometry method. Instead of CUDA, it uses the vendor-agnostic Vulkan API, achieving state-of-the-art accuracy with causal real-time estimates. It was deployed on a workstation, a laptop, and a very cheap single-board computer, outperforming CUDA-based systems on the same hardware — enabling low-latency, low-power VIO for robotics and XR.
More from Embodied
- Meta Connect hackathon: apps built in hours that once took teams weeks — HarperSCarroll · 2026-09-29
- VLS: training-free VLM steering of frozen robot policies wins CoRL and CVPR workshop awards — micoolcho · 2026-09-29
- Earth Rover Challenge: robots run policies in the wild in Costa Rica, dodging curious kids — micoolcho · 2026-09-29
- TrackEverything: 3D de-duplication breaks long-horizon dense point tracking trade-off — zhenjun_zhao · 2026-09-29
- AI robotic integrator agent handles end-to-end MicroFactory deployment — ihorbeaver · 2026-09-29
- Meta's Phoenix Glasses CAD Model Spotted in LinkedIn Video, Fans Want Open Files — pvncher · 2026-09-29