VR & Augmented Reality

Valeo Comfort And Driving Assistance received 1 patent for an in-car mixed reality gaming system that transforms vehicle sensor data into interactive experiences for passengers. The system takes inputs from automotive sensors, typically used for autonomous driving functions, and converts them into a game coordinate framework that maps real-world vehicle movement to avatar actions within location-aware games running on mobile devices. Each trip generates a unique game world tied to the actual route and surroundings captured by the car's sensor suite.

HTC received 1 patent addressing mixed reality headset performance through dynamic adjustment of pass-through camera image quality. The system monitors both GPU load and frame rate simultaneously, using these dual metrics to determine in real time how much quality to allocate to the video feed showing the real world through the headset's cameras. This treats the pass-through imagery itself as a tunable resource that can be dialed up or down to maintain overall system performance, rather than only adjusting the quality of rendered virtual objects.

Roblox received 1 patent for technology that synchronizes real user head movements with 3D avatar orientation in virtual environments using mobile device cameras. The approach combines device orientation sensors with facial landmark tracking to determine head position and rotation, then applies a configurable portion of detected movement changes to update both the avatar's facing direction and the background perspective across video frames. The tunable responsiveness prevents the avatar from overreacting to minor head adjustments while maintaining a natural connection between physical and virtual motion.

Niantic received 1 patent for validating the accuracy of AR mapping data before users attempt to start location-based games. The system performs simulated consistency checks by analyzing pairs of frames from validation scans, comparing how much the camera pose should have shifted based on sensor measurements against how much shift the AR map predicts between those same frames. This generates localizability scores for points of interest that indicate whether a given location will support reliable AR experiences, allowing problems to be identified before gameplay rather than during it.

Microsoft received 1 patent for adaptive frame rate control in head-mounted displays that adjusts refresh rates based on changes in the critical flicker fusion threshold. The system tracks how stimulus characteristics such as brightness, contrast, and color affect the minimum refresh rate needed for flicker-free perception, then modulates the display's frame rate accordingly in real time. This perceptual approach ties display hardware behavior directly to human visual biology rather than relying on fixed refresh targets or simple frame time measurements.

Apple received 1 patent for deep learning-based image reprojection designed specifically for AR and VR headsets. The system uses neural networks to predict and reconstruct frames between fully rendered images, reducing the computational load required to maintain smooth visuals during head movement. The models are causally constrained to avoid introducing artifacts from future frame data, addressing the latency sensitivity inherent in head-tracked displays where any delay between motion and visual update becomes immediately perceptible.