Speed
Pucks travel at 160kph. Audio cues are too slow. We use UDP Multicast for instant tactile reaction.
Valid Data
Open-Source • Assistive Tech • VI Hockey
Replacing Auditory Overload with Tactile Intuition.
An open-source hybrid tracking ecosystem for high-speed VI sports. < 50ms Latency. 100% Hacker Friendly.
Latency 50ms
TrackingValid Data
VisionComputer Vision
The Core Problem
Durability
Sensors inside pucks shatter. We use “God View” computer vision to track standard pucks from the rafters.
Computer Vision
Occlusion
Cameras lose sight in scrums. We fuse vision with UWB (Ultra-Wideband) radio tags so the player never disappears.
Occlusion/Prediction
The Architecture
A visually distinct tech stack engineered for speed and trust.
The God View
Dual-mode tracking using YOLOv8 for detection and “Blur Vector Analysis” (Hough Transform) to treat motion blur as velocity data.
The Fusion Engine
A Distributed Kalman Filter. It arbitrates between Camera data and UWB tags, using physics to “ghost” the puck’s location when visual contact is lost.
The Shear Glove
No vibration. We use lateral skin shear to create “phantom forces” that physically pull the hand toward the puck.
Feature Highlight: The Kalman Filter
Predict. Correct. Repeat. Our implementations of the Kalman Filter don’t just smooth data; they predict the future during visual occlusion, ensuring the haptic feedback never stutters.
Developer & Community
Built on Python, OpenCV, and Raspberry Pi. Join the contribution.
PythonOpenCVRaspberry PiUWB
{
"protocol": "OHS/1.0",
"puck": { "pos": [12.45, 8.12], "vel": [15.2, -2.1] },
"latency": "8ms"
}Ready to build accessible speed?
Review the system design, protocols, and haptics research.
Read the Whitepaper